Emerson has been working with BP, as we reported on Monday, to prototype their new SmartWireless offering. Here’s what they did at BP Cherry Point refinery in Blaine, Wash.
Science Experiment or Not, Wireless Worked at Cherry Point
Dan Carlson, senior wireless engineer from Emerson described the pioneering installation of Emerson’s self-organizing mesh network wireless sensor system at BP’s Cherry Point refinery in Blaine, WA as, “a science experiment, if you will, but over the next two years, we found it really works!”
Marty Gering, wireless data collection coordinator and wireless worker administrator for BP Cherry Point worked in construction, then joined the refinery and began working with the wireless data collection program—which made him ideal to be the one responsible for prototyping Emerson’s Smart Wireless system at Cherry Point.
Although Cherry Point is actually the second newest refinery in the US, there are many locations where large bodies of uncaptured data exist. “This data is valuable,” Gering said, “but we can’t touch it because of the expense of wiring and running conduit. Lube oil and bearing temperatures, among other values, are just left out of the picture.”
BP’s goal in partnering with Emerson in this product development exercise was to make it possible to do more process monitoring, more maintenance monitoring, and more environmental monitoring—all things that are generally considered too expensive to do in a wired world.
“So we began by automating operator rounds,” Carlson said. Gering chimed in, “This is one of the dirtiest parts of the plant, and the bearing boxes were covered with a layer of grease, oil and coke dust. The operators would use Raytek guns to take the temperature readings, but because of the environmental factors the handheld temperature guns weren’t all that accurate. So we installed wireless temperature sensors in the bearing boxes themselves.”
And that’s an oops!
“Emerson hooked up the network and got all the sensors talking to each other the night before they brought them over to the plant because they wanted to impress us, but they brought them over after they’d worked just fine in the hotel parking lot, and they didn’t work,” Gering said.
“We got about 5 feet of transmission,” Carlson explained, “and when we went around to find out why, we found an RFID reader broadcasting on exactly the same frequency. When we went to the channel hopping time synchronized system we have now, the problem went away, and we’re getting transmission distances of 300-plus feet.”
“You need some way of managing the RF sources in your plant,” Carlson noted, “without site surveys and complicated frequency analyzers.”
“This underscores the need to own your own wireless space,” Gering said. “We set up a plant wireless governance group to deal with this issue. I am on it, the I&C supervisor is on it, there’s somebody from IT on it, the guy who is in charge of our Maximo CMMS system is on it, and a couple of others. Basically, if you want to bring a wireless device onto the plant site, you have to bring it through us.”
If it looks like a duck…
“One of the things we were most fearful of was the acceptance or lack of it by the operators,” Gering noted. “But when Dan took a familiar blue Rosemount transmitter out of the box at the daily operators’ meeting in the I&C shop, it became obvious that the units looked like a wired device, commissioned like a wired device, and operated like a wired device.”
“It is all about managing cultural change,” Gering went on, “Once they saw the blue Rosemount housings, you could hear the techs relax.”
The wireless network doesn’t connect directly to the DCS. Instead, it goes via a 802.3 optical network to the OSIsoft PI server where the data is recorded in the PI database. “This gives us the flexibility,” said Gering, “to route the data where it needs to go. It might not be the best thing to load down the DCS with hundreds of bearing temperature readings.”
Gering and Carlson showed a short video clip of Gering himself going out with an intrinsically safe handheld and pulling a barcode into the handheld, and then pulling the wirelessly collected data from the PI server. “The PRIDE system permits us to access specific data from the PI server just by scanning a barcode,” Gering said.
Carlson said that they found that the existing network “acted as a huge antenna for expanding the network for easy scaling.” So they added vibration monitors to the network.
When asked if BP planned to use wireless in control, Gering said, “We’re sticking to monitoring and alerting only, but we’re designing the systems so that if we choose to move to control in the future, we can. So the answer is later, but not yet.
One of the concerns the refinery had was would they need to add new MOC (change management) items for tags, and so forth. As it turned out, the naming conventions, tags, calibration, and so forth are all standard.
So what did BP learn?
“I think it is good to start small,” Gering said. “I think it is cool that Emerson has that Starter Kit they announced this week is great. It is inexpensive and you can use it to build your confidence in the technology.”
The next project is the tank farm. Cherry Point was built to be a 95,000 bbl/day refinery, and has now become a 245,000 bbl/day refinery just by increasing throughput and optimization. The tank farm was built with only a level and sometimes a temperature sensor. “We’re beginning a new project to wirelessly connect those tanks with backup level. We also have hundreds of valves that we’d like to have positioner information on, as well as other pressures and temperatures,” Gering announced. “We have mixers with motors we’d like to monitor current on, and lots of other things. We’re looking at 300 points. We want to finish the engineering this year, and have a big start on implementation by the end of 2007.”