Manufacturing companies don't buy technology because it's trendy. In this industry, a technology had better be practical, affordable and carry a demonstrable return on investment, or they're not having any of it.
Wireless technology may be hip right now, but to manufacturers the ability to connect factory floor sensors without wires is anything but cool. A better word is practical. Previously, to network these sensors was either prohibitively expensive or flat-out impossible. A new crop of products and standards promises manufacturers the opportunity to do something they have never been able to do before: connect factory floor sensors on a network and interface those systems with broader enterprise networks based on Ethernet.
"These solutions allow manufacturers to extend their eyes and ears into the manufacturing process where they couldn't go before," says Tod Riedel, co-founder and vice president of business development for MillennialNet (Cambridge, MA), a low-power wireless sensor network provider. Wireless networking removes real-world limits such as the cost of cable, power-management hassles and the physical difficulties of getting to hard-to-reach equipment.
But the biggest benefit is getting a new window into what is happening in a manufacturer's processes. Rather than having an employee laboriously collect data from sensors once a day, for instance, these wireless networks enable the sampling of conditions every few seconds. The wireless sensors also provide a way to track and manage assets.
For example, to install a wireless sensor in an oil refinery, based on the ZigBee protocol, might cost $10 whereas installing wiring would cost $1,000. Having someone check that sensor once a day incurs high labor costs, inaccuracies and includes personal safety risks as well. Putting that sensor on a wireless network means you can poll it for data anytime or set it to automatically send data several times per minute. Formed in 2000, the ZigBee Alliance (the name hints at the way honeybees communicate) is working to ratify a low-cost, low-power protocol for wireless sensor networks. Ratification is expected near the end of the year.
A NEEDS-BASED SOLUTION
"We are out here solving needs, not catering to wants," says Venkat Bahl, vice chairman of the ZigBee Alliance and vice president of marketing for Ember Corp. (Boston), a ZigBee Alliance member. Ember sells wireless chips and software based on ZigBee to OEMs.
Another vendor in this space, Minneapolis-based Honeywell Automated Control Solutions (ACS), is working with the U.S. Department of Energy on a multi-node wireless pilot. "Our goal is to proliferate sensors in more industrial applications so companies can get more knowledge out of their systems," says Dan Sheflin, chief technology officer for Honeywell ACS. The company expects to launch its Honeywell XYR 5000 Wireless Transmitters in 2005 or early 2006. Honeywell ACS is a ZigBee Alliance member.
Sheflin sees a need for wireless monitoring in industries where harsh conditions can cause problems. For example, in chemical plants pipe erosion caused by abrasive chemicals is a continual issue. Though the damage caused by a chemical leak could be astronomical, generally chemical companies monitor for corrosion very infrequently since it requires removing the insulation and taking a picture of the pipe's condition.
With Honeywell XYR wireless technology, Sheflin envisions a company could place a series of corrosion sensors underneath the insulation on the outside of the pipe. These sensors could be programmed to send information as often as managers would like. "This is a much more thoughtful and cost effective way of doing maintenance," according to Sheflin.
Of course, like the ZigBee-based products, the Honeywell XYR sensors are not yet available. The promise remains just that-an expectation that this technology will be useful at some future date. "It doesn't make sense for a large manufacturer to beta-test new technologies on their established product lines. Wireless sensors are two years away," says Ian Barkin, managing director at The FocalPoint Group LLC, a consulting firm.
Barkin is not impressed with the ability to avoid laying wire. "There is Cat 5 cabling everywhere in most industrial settings," he says. What he sees as the strong potential for wireless lies in enterprise-wide asset monitoring and maintenance on an open protocol. "Traditionally, there was no hope for these things to communicate with each other," Barkin says.
These technologies are so new-the products are still in early development and most vendors in the space are still venture capital-financed-that it is tough to find manufacturing companies that can discuss beta experiences or even confirm they are interested in the technology. Early wireless networking adopters are likely to feel this is a competitive advantage they don't care to discuss in the open.
WHY ZIGBEE? WHY NOW?
ZigBee came about because vendors supplying the industrial and building automation space (Honeywell, Invensys and Philips were early members) were interested in leveraging wireless. But the existing wireless standards-WiFi (802.11), Bluetooth and Home RF-did not fit. The group needed a low-power, low-cost, small form factor, low data-rate solution. The Alliance partnered with the IEEE to define the physical and media access controller (MAC) layers of a lightweight wireless standard. This particular standard, which covers the wireless radio, was finished in April 2003 and is known as 802.15.4.
Now, ZigBee's focus is on defining and testing the network, security, and application software layers of the standard (these parts will be known as ZigBee), as well as getting the word out. Though final ratification is not expected until the end of the year, Bahl points to Ember's release of chips based on ZigBee as a sign that the standard is solid.
Meanwhile, testing is ongoing. "We don't just write the specs and put them out. If you don't test it, there are a million things that can go wrong," says Bahl.
Given that the standard is not yet formalized and actual wireless products are at least a year or two away, why should manufacturers care about ZigBee now? Bahl's answer is simple: "If you're not looking at wireless, your competitors are." He encourages manufacturers to get involved in the alliance, which currently has 80 members, and see how this technology might affect their plants. Member companies are the first to be involved in pilot projects that prove the concept of wireless sensor networks. "Focus on one wireless application today and start scaling your network tomorrow," according to Bahl.
And companies would not place themselves at risk by implementing today a sensor that uses ZigBee chips from Ember, says Bahl. "We have over-the-air upgrade capability. So if you deployed the sensors today, without replacing the chips you could upgrade the sensors wirelessly to make them ZigBee-compliant" when the standard is complete, according to Bahl, without the need to change out the hardware.
ZigBee in general-and Ember technology in particular-will be able to interface with other network protocols including TCP/IP running on Ethernet. "There is a very nice fit between ZigBee and Ethernet," says Bahl. This is important, since Ethernet is becoming pervasive in the industrial setting.
If he worked at a manufacturing company, analyst Barkin says he would hold off on investing in wireless sensors for the moment. "ZigBee would be a nice facilitator because it simplifies everything," he says. But there is no way to know in advance if it will take hold. "Companies will settle on ZigBee if no better alternative comes along."
Ultimately, wireless sensor networks are useful only if they enable companies to improve maintenance or better manage assets as a result of the information being sensed, communicated and acted upon. After all, investing in technology for technology's sake just isn't cool anymore.