Changes spurred by advanced computer modelling work on its regenerator are saving BP Bulwer Island Refinery millions of dollars each year in operating costs.
CSIRO Minerals used computational fluid dynamics (CFD) to diagnose and solve flow distribution problems in the Queensland plant’s fluidised bed catalytic cracker (FCC) regenerator.
And the work could be a world first. Dr Phil Schwarz, research program leader for fluids process modelling at CSIRO Minerals, says simpler modelling techniques have been applied to regenerators in the past.
“But they would not have been able to analyse the BP situation in the way that we did,” he says. “Nor would they have been able to diagnose the problem because they did not involve a complete, three-dimensional reactive model of the fluidised bed. This is the first time the regenerator of an FCC unit has been analysed in this way.”
Dr Schwarz and John Lee, BP Bulwer Island Refinery’s process support superintendent, won the Institution of engineers Australia John A Brodie Medal — awarded for contributions to chemical engineering — for their paper on the work presented at Chemeca 2006, an annual chemical engineering and processing conference.
The modelling project, which included coke combustion and heat transfer in the regenerator in addition to the complex hydrodynamics of the fluid bed, began after the refinery struck problems with its regenerator in 2000 following the installation of a new catalyst cooler.
Lee says that although the company had been involved in CFD work with CSIRO on another section of the FCC in the late 1990s, he originally discounted approaching Dr Schwarz for help, believing the modelling involved would be too complex to be completed in the timeframe required.
“It had to include reaction kinetics and temperature profiling on top of CFD that was already quite complex, being a gas fluidised bed,” Lee says.
However, the BP team, which believed it understood the problem and how to fix it, ultimately enlisted Dr Schwarz because of the high potential cost of getting it wrong.
Time was still a concern, though. At one stage of the research, Lee despaired over the fact it was taking about two weeks of computer time to replicate only a few seconds of real time in the regenerator — meaning the modelling work would take years.
But again CSIRO came through, with Dr Schwarz producing a technique that accelerated results. Further simulations confirmed the accuracy of this groundbreaking model, giving BP strong confidence in the outcome.
Work was already showing temperature differences from one side of the Bulwer Island regenerator to the other — an important discovery in a place where heat balance is critical.
The refinery uses a cracking process to break heavier hydrocarbons in its crude oil feedstock into lighter ones. Catalysts used to promote this cracking reaction are recycled — entering the regenerator, where coke deposited on the surface of catalyst particles is burnt off. This prepares them for re-use while also producing heat, which helps meet the energy needs of the cracking process.
As well as the temperature differences, Dr Schwarz’s work also found that combustion in the bed was incomplete, with some carbon monoxide leaving the bed and burning to carbon dioxide elsewhere, which is detrimental to energy efficiency.
He says heat balance is critical in the regenerator, which contains a fluidised bed of catalyst particles kept in motion by air bubbling up from below.
“If some areas have a high temperature and some a low temperature, you won’t be able to burn off the coke without increasing the heat to the critical temperature, where the effectiveness of the catalyst is severely degraded,” Dr Schwarz says.
Too low a temperature creates an inefficient combustion process, impairing energy recovery.
“We extended a CFD model describing the hydrodynamics in the fluid bed to look at the combustion processes occurring and the temperature distribution that would arise from those reactions.
“On analysing results from that model, it became clear that catalyst returning from the cooler was not being well mixed, and the air injected with it was short-circuiting through the bed.”
Dr Schwarz says BP personnel played a key role in putting forward potential modifications to solve the problem. “The CFD model assessed those potential modifications … allowing the team to decide which was the preferred modification.”
BP elected to change the arrangement, splitting the catalyst coming from the cooler and feeding it into opposite sides of the regenerator. The modification is saving the refinery several million dollars in annual operating costs.
“We’ve greatly improved the efficiency of the burn inside the regenerator,” Lee says. “It uses less air to burn the same amount of coke in there, which is usually the key constraint in running a catalytic cracker.
“It has also reduced the large temperature profile that we had within the regenerator.”
The modification has also helped increase the reliability of some regenerator components and, because less oxygen is being added to aid combustion, has reduced sulphur trioxide and nitrogen oxide emissions from the stack.
Phil Schwarz, CSIRO Minerals