System migration plan features on-the-fly control network split

Aug. 21, 2007
Expansion plans at Petro-Canada’s Edmonton refinery entailed a somewhat risky on-line split of its LCN to create additional network capacity in a way that ongoing operations were minimally affected.
At one of Tuesday’s “Lunch and Learn” sessions at HUG Americas 2007, Eric Lau of Petro-Canada and Sigurd Imgrund, project manager for Honeywell, told the story of a control network split and system upgrade at Petro-Canada’s Edmonton, Alberta, facility. A complex project at best, this one was complicated by the fact that Petro-Canada needed it done while the refinery was up and running and while other construction was taking place on the site.

The Edmonton refinery has been growing rapidly over the last two years and needed to expand its capacity. At the same time, Petro-Canada decided to convert its operations to 100% oil sands, or synthetic fuel refining. In addition, the company needed to build a foundation for further control system migration from Honeywell TPS to Experion PKS.

Petro-Canada’s existing control architecture included a main Local Control Network (LCN) controlling all refining process units, and a pumping and shipping LCN controlling crude and product tanks and pipelines. With the increased capacity and demands, the main LCN was simply overburdened, operating well beyond the recommend 55% normal and 75% peak loads much of the time, explained Lau. The system included 46,000 DCS tags and 1,300 graphics, as well as a number of customized programs and extensive cross-LCN integrations. It also had a high node count and insufficient capacity to support facilities expansion.

To alleviate this situation, Petro-Canada made the difficult—and somewhat risky—decision to do an on-line split of its LCN to create additional network capacity, segregate areas of operation and ensure reliable systems. “The decision was not a small deal,” said Lau. “There were a lot of control system issues.”

The project had to be implemented in such a way that ongoing operations of the refinery were minimally affected. The current operation was highly efficient and optimized, and management wanted as many of those characteristics retained as possible, as well as a one-stop control console for operations, plus interchangeability between areas of responsibility.

In order to accomplish this, a team from Honeywell, Petro-Canada and engineering and construction company Bantrell was assembled and tasked with splitting the LCN with the following strictures: The new system had to have capacity for future growth; the upgrade had to be done in a stable, reliable fashion and have no impact on current operations or any of the on-going construction at the refinery.

The operation was done in two phases late last year: First the team had to remove the utilities from the main LCN, append them to the performance management system and break the network gateway to allow certain network connections to take place. In the second phase, the synthetic crude production and the light and heavy oils production were put on separate LCNs.  

In preparation for the upgrade, the team built a test bed where procedures were tested and bugs worked out ahead of time. While this added to the expense of the project, it greatly smoothed the final upgrade.

After weeks of planning and testing, the actual project was completed in a 4-day window during a plant shutdown for other reasons. Lau had initially assumed that the upgrade would be done while the plant was running, but late in the planning phase, the team discovered that because of ongoing construction at the refinery, there would be a planned 6-day shutdown, so they adopted their own schedule to have it occur during that time.

Sigurd Imgrund, stressed the importance of organization on such projects. “Communication and organization were essential,” he said. “The 80-20 rule applies. It’s 80% organization and preplanning and 20% actually doing the work.”

The network split is part of a long-term plan at Petro-Canada. The next phase involves migrating the entire site to the Fault-Tolerant Ethernet network and then, with the upgraded infrastructure in place, bringing the same capability to the piping and shipping part of the operation.

“It was not an inexpensive operation,” said Lau, “but it was well worth the effort.”