THE CALL OF wireless in industrial process plants is seductive. No wires, no wiring, lower cost, easy replacement and retrofit… it seems like it’s too good to be true. Well, it is too good to be true. Yes, companies are selling wireless products in industrial process applications, but this is still the very early days, and the old saw about the dancing bear applies. You know, “it’s not that the bear dances well, but that it dances at all.”
Most of the applications currently in use are in inventory control or maintenance management systems, not process control. Nearly half the respondents to a recent CONTROL survey didn’t believe wireless was robust enough for process control applications. Of course, over half of the respondents said they believed that eventually it would be.
Startups, FUD and vaporware
Fueled by the return of venture capital to the technology markets for the first time since the telecoms and Internet bubbles burst in 2000, many companies are staking their claim to be “the leading player in wireless sensor technology.”
The buzz is biggest in building automation and home automation, because the numbers guys that work for the VCs can count the number of buildings and homes, and the potential number of sensors, and, as former Illinois Senator Everett Dirksen famously remarked, “Pretty soon, you’re getting into real money.” Even the U.S. Department of Defense and IBM are getting into the act. The DoD’s dream is to be able to drop sensors like dustmotes across a battlefield, and be able to “see” a soldier’s-eye-view. So it’s also dropping money into wireless sensors. If the technology development in wireless sensors mirrors the last big thing DoD got hot about (RFID), we should see some striking improvements.
But, for the past three years, we’ve been trying to get people to discuss their existing wireless applications, and we’ve had little success. Either they aren’t there, or they’re not working well, or there is much more secrecy involved with wireless projects than any other process automation projects we know about.
“We are going to see the early adopters, the marginal projects, for a while yet,” says Gerald Niemi, product manager for wireless at B&B Electronics. “Until standards settle down and products become fully industrially hardened, we won’t see widespread use. “This is still the Wild West.”
One of the most developed set of applications is for energy management in large buildings and campus settings. Nels Tyring, the cutting-edge systems integrator and chairman of Automation Alliance Group, pointed us to a company named Obvius with a wireless package aimed specifically at energy management systems, and several large-scale systems installed, including one at the University of Utah’s Salt Lake campus. ”They have been in the energy monitoring business for some time and have a very comprehensive package,” Tyring said.
Bandwidth and Propagation: The Physics of Radio
Unlike all the other analog and digital communications methods currently in use in industrial process automation, wireless is not clearly easy to apply. Whether you are communicating in 4-20 mADC current, or HART (digital superimposed on analog 4-20), or via one of the other industrial fieldbus protocols, or via Ethernet, the fact is, you have a wire going from point A to point B, and if the wire is intact, the signal automatically goes from A to B, by definition. There are certainly problems with signal attenuation, especially over long distances, but these are straightforward engineering problems with recognized solutions.
Contrast this to radio, where signal transmission is inherently problematic, and a function of power, gain, interference, shielding, and even sunspots. A wireless signal is broadcast, not narrowly transmitted down a wire. It can interfere with other signals, and be interfered with in the same way. Like all RF energy, it can be shielded against, either deliberately, or more often inadvertently, by geographical features, buildings, and interior structures. These include steel beams, stairways, catwalks, and process vessels, all of which abound in process automation applications.
David Kaufman, an SP100 member from Honeywell, says, “The operative issue is not how to make radio impervious to interference, but how to make error-checking and transmission protocols sufficiently robust, so expected signal losses don’t cause errors in process control or data transmission.”
Enter “mesh networking” (See Figure 1 below). Widely confused with one of the IEEE wireless standards, Zigbee, mesh networking is not a standard, nor is it a protocol. It’s actually a network topology. Mesh networks exist completely independent of radio, and chances are that you’ve used both of the largest existing mesh networks already today. They are, of course, the Internet and the telephone switching system. And if you are using VoIP for your telephone connection, you’re using both mesh networks at the same time.
FIGURE 1: MESHING WIRELESS NETWORKS
A mesh network is a set of router/retransmitters that form a self-organizing and self-repairing network.
The idea behind a wireless mesh network is that redundancy is the savior of data loss. As with the famous Internet protocol, TCP/IP, mesh networks route signals between short distances to router-redirectors in a self-organizing, and self-healing fashion. That’s the good news. The other piece of good news is that there is a standard: IEEE 802.15.4 or Zigbee, which uses the mesh network topology.