Regenerator head replacement project
How one of the world’s largest FCC regenerator heads with cyclones was removed and replaced during a 30-day turnaround
John Payne, Foster Wheeler Energy
Rob Tarbard and Peter Johnson, Esso Petroleum Company
Viewed : 7417
Head removals to change out reactor and regenerator internals are quite common, so what made the regenerator head replacement at Esso’s Fawley refinery in the UK different? The 17m head diameter, 20 pairs of cyclones and a total lift weight of approximately 700 tonnes, all to be achieved in a limited plot space, made the Fawley regenerator head replacement one of the largest head replacements ever carried out in the world (Figure 1).
By paying meticulous attention to detail, investing in pre-turnaround activities and making important decisions early, the Fawley refinery was able to execute its fastest ever fluidised catalytic cracker (FCC) turnaround. In fact, the 30-day turnaround duration was not dictated by the regenerator head replacement, but by other project and maintenance work.
The FCCU was originally site built in the 1950s, and the regenerator cyclones and plenum were last replaced in 1981. Due to limited crane capacities at that time, work was done piecemeal, with the cyclones removed and installed as pairs. This resulted in an extremely long turnaround.
From the previous 2001 turnaround, Esso had determined that, after 20 years’ operation, the cyclones were coming to the end of their life and any repairs in the planned 2006 turnaround would not be practical and would present a risk to future operational reliability.
In April 2003, approximately three-and-a-half years before the proposed shutdown date, a study was initiated to determine the best way to change out the cyclones, deliver the new head, plenum and cyclones, and then build them on-site, all within a 30-day turnaround.
The team quickly discounted the use of a large window in the side of the regenerator shell, as it would require a shutdown in excess of 60 days. Instead, the team decided that the only way to achieve this short turnaround was to install a new top head, complete with new cyclones. Two initial options for the head replacement were studied:
• A very large steel bespoke gantry system using strand jacks with sliding frames
• A very large crane.
The study assessed the following factors:
• Shutdown schedules
• Project schedules
• Risk management
• Head frame design concepts, with tonnage
• Piling and foundation requirements
• Build method for both crane and gantry
• Estimated lift weights
• Head cutting methods
• Cyclone layouts and budget costs from cyclone suppliers
• Concepts for plenum design and cyclone hanger system
• Identification of existing equipment to be relocated.
The gantry system was more costly, and there were a number of technical challenges relating to the build and operation that proved difficult to resolve. For these reasons, this option was not considered any further and the crane option was developed in detail.
Mammoet was awarded a contract to carry out some detailed conceptual studies to evaluate various crawler and ringer cranes, and how they could be built and operated within plot constraints.
Removal of existing plant
Considerable investment was required in relocating the existing plant, to accommodate the assembly and operation of the main crane and head frame assembly. A spent catalyst unloading hopper, fines skid and a section of pipe rack were dismantled from within the live operating plant and relocated. The pipe rack included a substantial amount of redundant piping, but also the main feed line to the FCCU.
New regenerator head
The FCCU at Fawley is in the middle of the refinery, making the delivery of large components difficult. At 17m in diameter, the new head could not be delivered to site in one piece and had to be designed to pass underneath the existing pipe racks en route to the final location that had a limited transportation height and width restriction of approximately six metres.
The new head (Figure 2) was designed to minimise site work by taking into account transport restrictions. The optimum design was found to be a head supplied in eight petal plates (the minimum number possible) and a central dollar section (Figure 3). The dollar section would be supplied complete with a plenum and refractory lining. The central plenum was designed sufficiently small to be supplied attached to the dollar plate. A 1.6m additional straight side of regenerator shell was required for the new cyclone design and this was supplied attached to the eight petal plates, again minimising site work.
Originally, the new head was specified as a 2:1 ellipsoidal design, but finite element analysis carried out by Foster Wheeler indicated that the shell thickness would have to be greater than 38mm to avoid distortion from internal pressure and possible damage to the refractory lining. At this thickness, the new head would have to be post-weld heat treated in accordance with the code. Additionally, the extra weight would have caused problems with the crane lift. So the team agreed that the new head design would be a hemispherical design, similar to the existing head.
Add your rating:
Current Rating: 2