Wireless wins in dire straits

Bryan McVicker transferred to ExxonMobil’s Beaumont, Tex., chemical plant in July 2008, but he got more than he had bargained for when Hurricane Ike tore through in September, leaving in its wake a broken levee, a plant flooded in 10 feet of water, and all electrical infrastructure destroyed. But after giving wireless a try (any port in a storm), McVicker and his team were sold on wireless benefits.
On Wednesday, at the Honeywell Users Group in Phoenix, McVicker, a Pyrolysis engineer at ExxonMobil, and Ken Routh, from the core engineering group at ExxonMobil Chemical in Baytown, Tex., gave their version of how they depended on wireless in this dire situation. “We were not allowed to go back to our offices or retrieve anything,” McVicker said. “It was all trashed. While employees worked out of hotel ballroom for several weeks, the entire organization had to restructure. When we returned to work, we found flares remained in service, but DCS and controls were lost. We were going to be in turnaround state for long time. We needed to make sure basic systems were in place to address potential hazards. I needed to make sure all core business units on site were adequately handling business preservation.” Having no previous experience with wireless, McVicker was hoping to supplement operator rounds, not replace them, as key process parameters during recovery.
Some of the challenges the company encountered included a weak infrastructure, no wiring, cabling, fiber optics, or permanent power. Also, they were limited in people because the instrument and electrician technicians and engineers were all focused on rebuilding the plant. So they contacted Honeywell to investigate wireless options. Within two days, the vendor had mobilized technicians, and wireless stations came to the site. Within 12 days, there was an initial startup with risk-based security and controls review and instruments limited to critical services.
As an expert called in to help, Routh’s role was to provide global support as a technical generalist. “Some of the biggest specs we had were quick response, and we wanted a fix that could provide everything,” Routh said. “We needed one vendor who provides everything from transmitters to operator interface. We needed 802.11 Ethernet backbone wireless because we had no fiber optic or cable. At least at that time, the vendor we could identify quickly was Honeywell.”
The Beaumont team implemented XYR 600 frequency hopping spread spectrum transmitters with five-second update rate. They initially installed nine transmitters, five-second update rate, with17 transmitters ultimately installed. The fix also included IEEE 802.11a wireless backbone mesh, four multinodes initially installed, with a final total of five, all powered with a solar panel (temporarily powered initially with generators).
With an operator shelter 1100 feet away from the transmitters, the team used just a couple of PCs, with one seeing use as a OneWireless server that did all configuration and wireless networking setup. “One of our security controls stipulation was this network had to be completely standalone,” Routh said. The only place to view the data was in front of a table in this operator shelter; there was no Wi-Fi capability. The interface was Honeywell HS Experion on a desktop PC.
“The initial configuration had two multinodes close together because of the tall structures in the plant and concern about interference,” McVicker said.
All in all, the wireless system made crucial device monitoring easier to use. It was staffed by console operators 24/7, with instrument monitoring and historization, flare pilot thermocouples, cold flare vaporizer TIs and LIs, and key utilities, such as firewater, potable water, natural gas, LP steam, and PIs.
Once the system was in service, operators immediately began using the system. They were able to troubleshoot systems with a proactive response to pre-alarms and alarms. At the end of the outage, data was collected. “The only thing we cared about was data appearing on the screen,” Routh said. “We decided we needed technical data toward the end, so the Honeywell folks in the Houston area spent a lot of time helping us configure and troubleshoot; we monitored seven key performance parameters for three weeks at five-second intervals.
In the end, there was no problem in the system with errors or retries. During this time, one of the multinodes failed, where there were two close together. “We were worried about transmitters being able to get to multinodes,” Routh said. “But we lost no data during that time.”
—Ellen Fussell Policastro