Online blend optimization boosts profits for Chevron

By Paul Studebaker

Jun 20, 2017

HUG Banner 2017

The challenges of implementing a sophisticated online analysis and real-time optimization strategy such as Honeywell’s Profit Blending and Movement (PBM) are not to be taken lightly, and the incentive to keep on using what already seems to work well enough may be strong, but the potential gains cannot be ignored. Especially if your existing system is facing major upgrades due to obsolescence.

“Some of us see lifecycle management as staying in the leading, but not the bleeding, edge,” said Paul Singh, blending specialist, Chevron, to attendees of his session, “Chevron Achieves New Benefits and Sustainability from Experion PBM Upgrades” at the Honeywell Users Group 2017 conference this week in San Antonio, Texas.

As a member of Chevron’s Business Improvement Network, which identifies emerging technologies and opportunities for using them to improve operations, Singh knows what he’s talking about. “I define control system lifecycle management as maximizing service life—operating the system as long as possible with assured support and future migration path—while staying at the leading edge to benefit from the latest technology, without affecting continuity of operation.”

A revolution in blending

At Chevron, blends are planned 15 to 30 days in advance to allow for material usage optimization. Multiple blends are planned, optimized and scheduled at the ERP level. Instructions are downloaded, and blend property control (BPC) is done on the process control network. Blend ratio control (BRC) takes place at the local level, based on readings from tank gauging, flow controls, knock engines, analyzers, etc.

In 2005, the company began a series of blend optimization implementations and upgrades. “We had blend optimization based on cost, but not on individual equipment constraints,” Singh said. Those constraints often limit the accuracy of traditional top-down blend optimization. “We needed to take into account the restraints on valves, gauges and other equipment,” Singh said.

Honeywell Profit Blend Control (PBC) optimizes using online constraints and downloads to the blending system. “PBC changes are small, since they’re done on top of the line profit optimizer—on the order of 1-2% of recipes,” Singh said. The refinery went from BPC 54.0 to OpenBPC Blending and Movement Automation (BMA) 320.7 in 2005, and to Profit Blend Optimizer (PBO) Profit Blending and Movement (PBM) 431+ in 2015. In parallel, the control side migrated from Blend Ratio Control (BRC) 44.3 to Experion Blend Control (EBC) BMA 320.7 in 2010, then to Profit Blend Control (PBC) PBM 431+ in 2015.

On the analysis side, in about 2008, the refinery shifted from near-infrared (NIR) analysis and knock engines to modeling based on off-line Raman analysis. “In 2011, we went to online Raman analyzers that give us octane, Reid vapor pressue—all the same data as the lab,” Singh said. “We just need one ¼-in. hole in the blend pipe, and we have 35 properties available at one-second intervals. In practice, we do some smoothing and averaging, take readings every three minutes and optimize every eight minutes.”

Motivation and ROI

Upgrading the blend control system was driven by concerns about safety, environment, reliability, security, standardization and Tier 3 requirements, as well as economic benefits. Safety pays off in fewer reblends—fewer starts and stops, less fatigued operators, and fewer safety incidents or near-incidents. “We used to have fairly high off-spec and reblending, about 13%. Now it’s about one reblend per month,” Singh said.

Increased blend accuracy means less giveaway of octane, expensive components and additives. “Based on our experience, you have the potential to save 80 cents per barrel,” Singh said. For a 50 million barrel-per-year refinery, that’s $40 million per year. “It may take a few years to get there—you can expect maybe one-fourth of that the first year—but our giveaway decreased by about 50%.”

Considering the upside profit potential, it’s not surprising that it’s not an inexpensive project. “It’s a highly customized and engineered application. Half the total cost is in PBM engineering,” Singh said. Don’t expect to get the same results from a generic optimization system.

The project execution was done in two phases. In phase 1, BPC was migrated to PBO and involved adding a PBO server with a 64-bit Windows Server 2008 R2 operating system and SQL Server 2012, redundant Experion servers with LCNP cards to communicate with the TotalPlant Solution system, and an Experion station.

Phase 2 completed the migration to PBM 431+. It added redundant Movement Automation Control (MAC) servers running 64-bit Windows Server 2008 R2 and SQL Server 2012, redundant C300 controllers with C300 I/O, and additional Experion stations.

PBM 431 features include support for redundant level and temperature sensors as well as application white listing. “Use wireless to back up the wired system and provide redundancy,” Singh recommended. PBC enhancements include supplemental pressure control by adjusting target flow rate; supporting the blend from startup to custom interlocks; and supporting up to 40 components and 20 additives.

The upgrade is compatible with Experion PKS 431, and lab data is integrated via PKS. It downloads lab data from LIMS via XML, and makes it available for review, approval and updates. It integrates and distributes approved lab results, online analyzer readings, and BPC model predictions for improved blending and movement management. Singh said, “On very large blends that may take more than a day, we can take mid-blend samples to the lab to check and improve the results.”

Singh concluded with a list of recommended best practices for refineries contemplating a similar blending upgrade:

  • Perform front-end loading study to accurately define project scope, schedule and cost in the face of fast-changing hardware/software options and lifecycles.
  • Employ virtualization where appropriate.
  • Get advanced process control staff involved in the project execution at the earliest stages of the project. “They ask a lot of questions, and they understand what it takes to get 99.9% uptime from the new system,” Singh said.
  • Resolve all instrumentation, electrical, mechanical and analyzer issues before system cutover.
  • Avoid cutover during the winter to summer transition. “Operators are already stressed,” Singh said. “Don’t add to it.”
  • Provide hands-on training and 24/7 support to operators until new system uptime meets expectations.
  • Test all interfaces and blend laws thoroughly, during both factory acceptance test and site acceptance test.

For justification, note that safety and environmental concerns motivate reliability and security, which leads to blend control optimization, on-spec products and return on investment. “Blending is the cash register,” Singh said. “It’s where custody is transferred and the refinery measures profitability.”

Future plans include multi-site, multiphase implementation of PBM-PKS R500+ Profit Blend Optimizer, Profit Blend Controller, Blending Instructions, Profit Inventory Monitor, and Profit Movement Management, including Movement Monitor, Movement Control and Tank Quality Integration.