Infrastructure options for industrial communication networks are diverse and often complex. When they don’t deliver the necessary degree of reliability, the process becomes jeopardized and unplanned downtime, or worse, can result. Most of the bad practices seen today relate to the components selected, installation practices or design. But with a solid architecture plan, the system will perform as intended.
The challenge is that industrial networks aren’t typically built from the ground up, but rather develop incrementally over the life of a plant. Plant expansions, relocations, technology updates and remediation are among the primary reasons for altering a network infrastructure. Or the goal might be strategic: bridging the manufacturing and corporate networks, strengthening reliability and security, or adopting predictive maintenance practices that require remote monitoring of asset conditions. Less common is the construction of greenfield plants, where an entire network is built from the ground up.
Hot topics: Ethernet and wireless
Open industrial network protocols like ControlNet and Modbus have been losing favor to Ethernet, which is extremely stable and has better speed, bandwidth, flexibility and communication management than the earlier protocols. Assembly lines and warehouse picking lines are often among the first areas to be automated with Ethernet. “Some areas of a network will remain on a ControlNet-like infrastructure, with its islands of information and slower data rates, when the equipment doesn’t allow for Ethernet connectivity or it’s not cost-effective to change,” says Mara White, marketing manger at Fluke Networks.
Likewise, networks primarily remain wired where high speed or safety is critical, but other hardwired applications are giving way to wireless alternatives using Ethernet bridges, Wi-Fi or cellular networks. Process I/O used in alarming, alerting and condition monitoring is increasingly moving over to wireless, for example.
Transitioning from wired to wireless is a selective process. “As the installation cost to wire becomes prohibitive or the capacity to install additional fiber or cables is insufficient, wireless becomes a viable alternative for these functions,” explains Michael Martinez from Invensys Operations Management’s Critical Infrastructure and Security Practice. “Once a wireless infrastructure is in place, incremental wireless solutions often follow,” adds Hesh Kagan, director of technology innovation for Invensys. “For instance, real-time location systems are very inexpensive to add on and deliver a large return on investment.”
A more ubiquitous wireless option might be needed to track assets that move among multiple locations, contractors and suppliers over large geographical areas. For those applications, cellular networks, such as AT&T Mobility’s manufacturing enterprise business solutions, fill the niche. Although the capacity of cellular networks is far lower than the alternatives, it’s suitable for many low-volume wireless asset management applications. Additionally, using a carrier-owned and -operated cellular network frees the end user from network maintenance.
Planning process: The logical view
When correctly planned, a network infrastructure will support its application and be productive. The first step is a detailed assessment of the network’s present state, from the logical, physical and security perspectives. What are its strengths and weaknesses? How does the network compare to industry standards?
“Most of the problems we see have to do with the network being built without a plan. It might have the wrong architecture or incorrect components or switches that don’t segregate traffic properly, and the network doesn’t work anymore,” explains Pete Morrell, global manager of network and security services for Rockwell Automation. In the planning phase, several considerations must be taken into account.
Application: The type of application determines the type of product used. What will it tie into? How far back into the corporate infrastructure? Users get frustrated quickly if the technology doesn’t fit the application. For example, an in-plant crane requires very different technology than a remote monitor at an isolated site.
Reliability: Wired networks are necessary when you need relentless speed and reliability. Wireless is suited to applications that don’t change very fast, like tank level monitoring. “In a 10,000-gallon storage tank, level doesn't change very fast, so any intermittent data loss because of interference doesn’t cause an upset in the process,” says Scott Killian, director of solutions connectivity for Sixnet. “In comparison, a high-speed machine application that requires lots of continuous data would not be the first location to try wireless.”
Practicality: Wireless provides the added benefit of convenience and, in some cases, can improve both uptime and reliability, such as those where cable is particularly subject to degradation. “Customers deploying wireless for the first time should ask, ‘Does it work for me?’ If they can’t get sufficient productivity and system performance for the application, the answer is no,” reasons Jim Weikert, strategic product manager for wireless at ProSoft Technology.
Environment: Basic cabling is copper, but if there is higher noise, fiberoptic cable is better. Sealed connectivity and abrasion protection are needed in extreme temperature, vibration, chemical and washdown environments. Noise and interference can be challenges for wireless in industrial environments, but application engineers now have many creative solutions. If you’re considering cellular wireless, check the signal strength. “As long as there is some type of 2G or 3G data coverage on a phone, then you know a data modem will work,” says Killian.