Nov-2020

Modelling ensures a successful revamp

Pre-turnaround computational modelling built confidence in the benefits of introducing design changes to a FCC regenerator revamp.

RAJ SINGH, PAUL MARCHANT and STEVE SHIMODA, TechnipFMC Process Technology
MARC A SECRETAN, Suncor Energy

Article Summary

For FCC operation, regenerator performance is a key factor in maximising unit economics. Regenerator performance is generally evaluated by uniformity of combustion, seen in the temperature profile, and strongly depends on good distribution of both air and spent catalyst. Both operating conditions and hardware configuration influence the air catalyst mixing and flow patterns within the regenerator.¹

Understanding the impact of hardware configuration on regenerator fluidised bed hydrodynamics is important for any potential design modification, optimisation, and troubleshooting. It is important to pre-evaluate the performance of the unit with planned design modifications to reduce any unforeseen risks before the implementation. TechnipFMC actively uses computational fluid dynamics (CFD) tools for design validation and troubleshooting.

This article discusses an FCC regenerator revamp at the Suncor Edmonton refinery, aimed to incorporate advanced design features to mitigate operational challenges and improve the mechanical reliability of the internals. This article describes how CFD modelling tools were used to confirm the adequacy of the proposed hardware changes, fine-tune the design, and evaluate the performance to minimise risks on start-up. Post-turnaround performance of the unit confirms the benefit of the implemented design.

Suncor’s Edmonton refinery is one of four refineries which Suncor operates in North America. The refinery was built in the 1950s, can process 142000 b/d of crude, with FCC throughput of roughly 45000 b/d. The FCC unit is a ‘side by side’ reactor/regenerator configuration which includes a close coupled riser termination device in the reactor and a regenerator with a fast burn zone. The layout of the FCC unit with its latest design features is shown in Figure 1. The unit has been revamped with multiple FCC Alliance (TechnipFMC, Axens, IFPEN, and Total) technology features in the last 20 years.

In 2001, FCC Alliance proprietary feed injectors and Suncor and TechnipFMC’s jointly developed riser termination device were installed in the riser reactor section. Post-revamp, the unit performance was significantly enhanced due to improved feed atomisation and reduced post riser residence time. Even though the regenerator operation was switched from partial combustion to complete combustion, the unit managed to achieve higher catalyst to oil ratio and lower delta coke. The unit experienced improved yields; gasoline increased by approximately 5 vol% and coke decreased by 1 wt%.

In 2004, FCC Alliance’s proprietary structured packing was installed in the FCC stripper. It was deemed a successful revamp as the unit was able to reduce the stripping steam consumption from 4lb to 2lb/1000lb catalyst circulation at a constant regenerator temperature. Since 2004, Suncor has inspected the packing during scheduled turnarounds and found no significant damage or erosion. It continues to perform effectively. Inspection pictures after 5, 10 and 14 years of service (see Figure 2) confirm that the packing is in good condition and continues to be reused.

Suncor’s FCC regenerator is a single stage, full burn regenerator with a fast burn zone in the lower section of the regenerator. It operates at higher superficial velocities than typical regenerators. The original internals consisted of a single horizontal arm spent catalyst distributor, ‘cross type’ air grid distributor, and an internal hopper feeding the regenerated catalyst standpipe (RCSP). The regenerator was experiencing an afterburn of 40°F (22°C) which infringed on the turbo expander inlet temperature limit. Additionally, high gas entrainment into the RCSP led to poor head build-up and low regenerated catalyst slide valve (RCSV) pressure drop.

For the 2018 turnaround, as the regenerator was determined to be at end of life, Suncor decided to replace the entire vessel and internals such as the air grid, cyclones, spent catalyst distributor, and so on. The original objective was to be a replacement-in-kind revamp. However, Suncor took the opportunity to incorporate some design improvements to the regenerator internals to mitigate the known operational problems and improve the mechanical reliability of the internals. The regenerator shell design was kept unchanged to minimise impact on the foundation, structure, piping, and external work required during the turnaround. The existing top hemispherical head was replaced with an elliptical head to raise cyclone inlet elevation, while keeping the total regenerator height constant.
 
Technology and design development
During the planning stage, technology enhancements focused primarily on spent catalyst distribution and RCSP operation. A self-aerated, submerged compound angle spent catalyst distributor was designed to improve spent catalyst mixing in the bed, promote uniform bed combustion, and reduce afterburn. The RCSP inlet was modified to reduce gas entrainment, increase catalyst density, and improve head build-up above the RCSV. A comparison of the original regenerator configuration with the new/modified configuration, with changes highlighted in red, is shown in Figure 3.

The modified configuration in Figure 3 shows an improved version of FCC Alliance’s standard compound angle wye bathtub distributor design, where a major portion of the distributor arm is submerged in the catalyst bed. Considering the regenerator diameter was small and the standpipe inlet was tangential to the vessel, a one-arm bathtub instead of wye arms was designed for this application. The initial angle of the bathtub is optimised to ensure incoming spent catalyst has sufficient momentum to flow down the arm, without aeration, into the fluidised catalyst bed. The latter portion of the bathtub is designed to self-aerate, utilising the gas from the bed underneath, and distribute catalyst preferentially to the centre and sides of the vessel. The design provides self-fluidisation and eliminates the need for a sparger system to fluidise the catalyst in the distributor. The design allows the catalyst to discharge in the high velocity lower section of the regenerator, in close proximity to the primary cyclone dipleg discharge. This promotes spent catalyst interaction with hot catalyst returning from the primary diplegs and overall mixing in the bed, which is essential to enhance bed combustion uniformity and reduce afterburn.

Since the revamp was originally intended to replace the regenerator and its internals in kind, there was concern that the proposed hardware changes may adversely impact the performance of the unit. To gain confidence in the likelihood of success and moreover to reduce any unforeseen risks, it was decided to evaluate and optimise the proposed design using computational fluid dynamics (CFD).

Computational modelling plays an increasingly important role in understanding gas particle flow dynamics in the FCC process, enabling designers to offer low risk, high value improvements. The latest generation of CFD modelling tools enables rapid exploration of different configurations to optimise the design. Compared to cold flow testing, CFD allows a deeper understanding of what is happening at all points in the system. It provides the qualitative information needed to visualise processes difficult to see in physical models, along with the quantitative results that enable realistic comparisons of equipment configurations. Combining these results provides an understanding of how hardware and operational changes impact gas catalyst flow dynamics in the fluidised bed. Conducting CFD ‘virtual testing’ of new devices, combined with experience, increases confidence in proposed changes.
 
Computational modelling
A CFD model was developed for the regenerator configuration with proposed internal modifications to study the impact on the fluidised bed hydrodynamics. Barracuda VR, CFD software developed exclusively to model gas solids fluidised bed reactors, was used for the study. The software was selected because of its capability to accurately handle dense gas solid flows. The algorithms and models incorporated into the Barracuda VR software have been validated against cold flow experimental data and commercial operating reactors.^2,3