1 April 2001

Ethernet Fends off Downtime

by George Reed

Textile plant fiber-optic upgrades eliminate coaxial transceiver breakdowns.

Like everyone, plant engineers love to sleep through the night. Unfortunately for engineers at Reemay's spun bonded textile manufacturing plant, Murphy's Law often raised them from their slumber.

Whenever one of the plant's coaxial cable network transceivers would go down, it always seemed to happen either in the middle of the night or when the plant was changing over a product line and needed the network. In some cases, several hours could go by before the problem was diagnosed and fixed. At a cost of $15,000 per hour for downtime, this situation was unacceptable.

To fix the problem, Reemay installed a redundant fiber-optic Ethernet network. Now, even if one of the fiber lines breaks, the system will keep working.

Networking Textiles

Located in Old Hickory, Tenn., Reemay manufactures two families of nonwoven spun bound products: Reemay polyester and Typar/Tekton polypropylene. The products are used in many applications, including primary and secondary carpet backing, furniture and bedding construction, fabric softener sheets, apparel interlinings, filtration products, automotive interiors, agricultural crop covers, landscape fabrics, geotextiles, and air filtration barriers. They also are used in materials for aerospace, industrial, and construction purposes.

One product is used in almost all fabric softener sheets. It is shipped to the various manufacturers who add fragrances and additives. Other products are shipped to Central and South America, Israel, and South Africa on a regular basis.

Reemay's quality management system includes stringent ingredient control; standard operating procedures and conditions for all critical plant functions; and physical property control using statistical techniques and standardized test methods such as TAPPI and American Society of Testing and Materials (ASTM).

Several years ago, Reemay installed Siemens PLCs to control the various machines involved in the manufacturing process, linking all the machines via a coaxial cable network based on transceivers. The network was state-of-the-art at the time of installation; but, as the components aged, they began to fail on a regular basis. If a transceiver failed, the entire network would go down.

When the network went down, it caused major problems. Although the individual machines could continue to run under local PLC control, workers could not make any changes to the machine or the line. For example, if operators wanted to change to a different product or material, they needed the network. One or more production lines would remain idle until the network could be repaired.

Repairing the network was not easy. The obsolete transceivers became hard to find and were becoming more expensive every time they were purchased. When it was later discovered only one manufacturer still made the transceivers, Reemay knew that prices would be going up substantially.

Management made the decision to replace the aging network with a modern system. The Old Hickory plant could not go to unknown and untested equipment, however, so research was done to identify known industrial standards with known equipment that would work. The modern system could not be proprietary because of the multi-vendor environment of devices, machines, and PLCs. Reemay wanted a system that could be expanded later. All these requirements narrowed down the choice to an industrial Ethernet system.

Networks Redundant

The network is set up in a dual ring structure, running from machine to machine around the factory floor in both directions. The dual ring structure provides redundancy and eliminates old "single point of failure" problems. If a machine, network component, or fiber cable fails, the system continues to function.

Siemens OSM (Optical Switch Module) and ORM (Optical Redundancy Manager) modules form a redundant network that operates at 100 Mbits/sec over a fiber-optic backbone. Each OSM is a six-port switch with two 100-Mbits/sec fiber-optic ports and four 10-Mbits/sec twisted wire ports. Terminal devices or industrial Ethernet segments connect to each OSM via industrial twisted wire pairs.

The network can operate in our electrically noisy textile mill. The galvanically isolated network is immune to electromagnetic noise per EN 50082-2, requires no lightning protection, and works at distances up to 3,300 yards without repeaters. The 100-Mbit/sec backbone can transport large amounts of data.

Adding in a new machine or device is simple. Systems engineers need only break into the system. An e-mailed message is automatically sent to inform the administrator that the ring is broken, but all other nodes are still working. A new device is wired to the nearest OSM, or a new OSM is added to the ring, and then everything is plugged back together. The system then sends an e-mail message that the ring is back up and running. Any necessary reconfiguration can be done from a remote PC. Expansion is virtually unlimited.

Reemay decided not to use a commercial Ethernet network, because none offered a high-speed redundant network, and they use slower protocols that are unsuitable for industrial applications. Other industrial networks were looked at, but none had a remote diagnosis and configuration capability of Industrial Ethernet.

Instead of being able to check conditions and set configurations from a PC, as is done with the Siemens system, engineers would have had to go out to the field and plug in test equipment. With the Web-based network management software that came with the switches, setup and configuration was easy. The network management software can be accessed remotely by dialing into Reemay's Intranet to monitor performance and allowing the network itself to remain secure.

Everybody claims their products are "plug and play"—but they rarely are. It was surprising to see that the Siemens equipment really did connect together simply.

The OSM modules allow network traffic to be re-routed into dedicated 10-Mbit/sec segments. This feature will be used in the future to eliminate useless traffic from the network. Today, the network at Old Hickory has peaks that reach up to 98% of load capacity without shutting down. The rule of thumb in the networking business is anything above 40% can shut the network down. Reemay will be using the OSM switches to reroute useless traffic off the main network, divide the one large network into smaller networks, and let the PLCs run at high speeds without problems.

Since the Industrial Ethernet system has been installed, unexpected, late-night visits to the plant are a thing of the past.

Savings up to $120,000/yr

The virtual end of unexpected network failures saves Reemany an estimated $100,000 to $120,000 per year.

Because the Industrial Ethernet system is Web-based and e-mail compliant, plant personnel know when there is a problem—before it affects the production line. The added ability to be proactive instead of reactive toward maintenance problems, and fixing equipment before the pending problem has a chance to get serious, saves an estimated $10,000 per year.

Finally, maintenance costs have been minimized. If a failure occurs in the middle of the night, the redundant network keeps on working. Reemay no longer has to call in a technician or pay overtime to troubleshoot the system. In most cases, problems can be fixed during normal business hours.

With the previous network, Reemay had to keep many spares on hand. If the system went down, enough parts were needed to cover every possible emergency, so its spare parts inventory was very large. Today, Reemay stocks only a minimal number of spare parts and components. If any part of the system fails, and a spare is not available, it will arrive by overnight mail. Meanwhile, the redundant system just keeps running—working around the failure. IC