Jan-2019

Analysis, active filtration unlock guard vessel stability (ERTC)

For the first time in years, they did not need jackhammers to unload the guard vessel. It was another short cycle, and there was another crust layer.

John Burwell
Crystaphase

Article Summary

But this was only the first load with Crystaphase, and already they had run 14 months instead of 12. With a 16% run-time extension, their best yet, engineers at this major North American west coast refinery had reason to believe they were finally on to something.

Six months earlier, a Crystaphase foulant lab technician had examined samples from the previous vendor’s system. Energy-dispersive x-ray spectroscopy showed high levels of carbon, consistent with the unit’s coker feedstock. But scanning electron microscopy revealed something more. The technician saw a rough, rocky surface on a dark mass of foulant that was fractured like a dry lake bed, crisscrossed with smooth-walled crevices. Crystaphase recognised the morphology. It was gum polymerisation.

Culprit identified, the team had their project: not just to spread out the crust layer, but to prevent it altogether.

Before calling Crystaphase, engineers had tried everything the catalyst vendor recommended. Each new proposal seemed to be a completely different configuration of the same basic materials, with no clear reason for the changes. Whatever they loaded, the unit would come down on pressure drop after less than a year, with a thick crust backed up into the poison control layer. Jackhammers became part of the unloading routine.

The frequent shutdowns kept the guard vessel out of sync with the naphtha hydrotreater (NHT) it served. That unit – among  the most valuable in the complex – was designed to run for four years. Turning around the guard vessel was a time-consuming and expensive nuisance by itself, but every shutdown of the guard vessel also disrupted the NHT's production, and every minute the NHT was offline was a minute the NHT was not making money.

And the NHT was hungry. Reducing feed rates at the guard vessel might lengthen cycles, but it would also cut volume, which helped nobody. Smoother feedstock would be an easy fix, too, but the refinery needed to process the coker feed. ‘Easy’ was no option.

Engineers and managers alike were ready to find a solution.

One engineer had worked with Crystaphase before, on that hungry NHT. There, the team had installed a system composed of ActiPhase and CatTrap, and performance data was thoroughly validating its technical advantages. The engineer was confident the NHT would reach its target 48-month cycle.

His next conversation with Crystaphase was about the guard vessel.

Following their usual methodical procedures, Crystaphase analysed samples of the previous grading material and proposed a replacement system: ActiPhase, CatTrap, and poison control. The idea was to move the reaction causing the crust layer to the top of the vessel, where the resulting particles would be filtered before they could form a crust. ActiPhase would trigger polymer-based particle generation, while both it and the CatTrap layers beneath it would capture the particles.

This was the system they were now unloading.

A Crystaphase technical service representative was on site to examine samples as they were collected. The objective was to determine which layers trapped more foulant, then expand those layers in the system they were about to install.

The ActiPhase layer was fullest. On-site calculations put it at 100% capacity. Each ActiPhase disc was solid and black like a hockey puck, indicating complete saturation. Yet no crust layer had formed; video feeds showed catalyst handlers moving pieces easily by hand.

Next, they unloaded the CatTrap. Again, the team calculated utilisation: it was less than 50%. CatTrap was working, but the foulant load was clearly higher in the ActiPhase layer.

If the ActiPhase layer was full because activity was a factor, then the poison control layers might also be full. That is when the catalyst handlers readied the jackhammers, only to realise they were not needed.

The crust layer, smaller than ever, came out with hand tools.

The principle was proven: active filtration could drop out the gum polymers. Now they just needed enough holding capacity in the upper layers to keep the foulant away from the poison control.

Crystaphase and the engineers quickly revised the system about to be installed. They had delivered extra ActiPhase. They would load all of it. The CatTrap they would reduce, and the poison control they would keep the same.

They wanted to load even more ActiPhase, but that would have to wait until the next change-out. In the meantime, they would take samples back to the lab for full analysis and use the findings to calculate exactly how much ActiPhase they would need for the following cycle.

At the end of this year, the unit will turn around again after running for almost 16 months – another small but significant incremental extension. In that time, Crystaphase and the customer have worked together on the next solution, based on observations from the past two turnarounds.
Analysis revealed iron sulphides in addition to the gum polymers. These often appear as products of the heating train, which remains in service, so the team planned to see them again.

Crystaphase measured foulant concentrations, and the customer established the ideal feed rate for the guard vessel. With these figures, they calculated the volume of foulant that would accumulate over their target cycle length of 24 months. They based the required volume of ActiPhase on that calculated volume of foulant.

The new recommendation calls for twice as much ActiPhase, in both 30 and 50 ppi sizes, to trap the gum polymers. To trap the remaining foulants, smaller layers of CatTrap in smaller pore sizes will go underneath. With the balance shifted toward ActiPhase, sufficient activity to drop out the gums, and sufficient capacity to hold 24 months worth of foulant, the customer will be set to run its longest cycle yet.

If it works as the customer and Crystaphase expect, they already know their next goal: 48-month cycles, finally putting the guard vessel in sync with the NHT. By the time Crystaphase tells this story again, the jackhammers may be long forgotten.

Did you know Crystaphase uses a replicable, methodical approach to extend cycle length?

This short article originally appeared in the 2018 ERTC Newspaper, produced by PTQ / DigitalRefining.

For more information contact: john.burwell@crystaphase.com