Uncovering the Problems Hiding in Your Control Loops

Sometimes the problems that limit production in some way are not readily apparent. I recently connected with Emerson’s Mark Coughran about his new role in consulting with oil & gas producers on ways to optimize their production. He and other industry consultants assist on startups as well as perform troubleshooting on unit performance limitations and implement advanced controls.

In most cases, oil & gas producers typically assign Mark and his fellow consultants to work on control loops that are causing problems such as throttled production, cycling, and spurious trips. Mark notes that their work is sometimes called “controller tuning” but it often evolves into fixing problems outside the process control system controller, into areas such as control valve performance, process transmitter behavior, and problems with PLCs on skid-based systems communicating with the control system. He shared four optimization examples with me.

The first was a nitrogen purification skid controlled by a PLC with all the variables displayed in a DeltaV system. The operators could see large cycles in the liquid levels, gas temperatures, and turbine speed. In this state, the startup team could not meet the schedule for turning the unit over to Operations. Mark and the team found that the levels and temperatures had slow integrating process dynamics as expected. This would normally call for large integral times in the PID controllers. The parameter values in the controllers looked OK until they discovered they were expressed in milliseconds, not seconds! By working with the PLC supplier, they rapidly stabilized these loops.

Steam Flow vs. Time on Regenerator/Reboiler Control LoopThe second example involved a steam flow to regenerator/reboiler in an acid gas removal unit. The steam flow is a slave loop to the temperature, which is critical to the properties of the regenerated solvent. The flow was cycling 10% of span in cascade (CAS) mode. Using the Emerson tools and years of experience, Mark and the team quickly eliminated controller tuning and transmitter performance as possible root causes. The evidence pointed to a problem in the valve positioner feedback linkage. A brief trip to the field with the instrument department confirmed this by visual observation. They corrected the problem during the night shift.

The third example concerned a water treatment unit in an oil sands steam-assisted gravity drainage (SAGD) ‎plant. The lime softener vessel experienced poor control of water level. This was due to a dead time of 200 seconds in the level response whenever a disturbance occurred with the flow valves fully open. A joint investigation with the process engineering team showed the valves were severely oversized for the application. The key recommendations for the plant then had nothing to do with the level controller PID tuning. Instead, it was to replace the control valves or to reduce the actuator travel, and to implement gain scheduling in the flow controllers.

Mark’s final example was a sour oil surge drum low level causing repeated trips of the dehydration unit. Again, the team found the problem was not inside the DeltaV PID controller. Instead, the problem was in the slave flow loop. The flow transmitter had been configured with a low-flow cutoff that stopped the flow controller from closing the valve when there was still significant flow out of the vessel.

As Mark shares in these examples, the problems are sometimes not readily apparent and require some sleuthing. The focus of Mark and the industry consultants is to work with your team to find the problems that limit your profitability, wherever they may be hiding in your control loops.

MP3 | iTunes


Leave a Reply