Jan-2010

Are there alternatives to an expensive overhaul of a bottlenecked flare system?

Changes in regulation or the revamp of an existing plant need not require an upgrade to a refinery’s flare system configuration

Detlef Gruber and Dietmar-Uwe Leipnitz, BP Lingen Refinery
Prasad Sethuraman, BP Refining Technology
Miquel Angel Alos, Jose Maria Nougues and Michael Brodkorb
Inprocess Technology and Consulting

Article Summary

There are various situations in which it is necessary to re-evaluate the capacity of a site’s existing flare system. In general, re-evaluation follows a potential increase in flare load; for instance, when:
• Relief valves that currently blow to air need to be connected to the flare header
• Changes in regulation redefine the scenarios for which the flare system has to be designed
• A new plant is connected to the existing flare header
• An increase in throughput or a revamp of an existing plant might increase the flare load through higher hold-up or higher heat load.

There are other situations that 
are unique to specific sites, but 
all of them seem to have only one solution when the capacity of the site’s flare system is reaching its limit: a capital project to increase its capacity.

Nevertheless, an effective altern-ative is available before an increase in the site’s flare capacity is considered. This alternative is based on the fact that most flare systems are designed on over-conservative assumptions and steady-state calculations for determining the flare load from the process units. Therefore, these flare systems might show additional capacity when analysed with a more adequate method that considers the dynamic effects of any release to the flare system. This alternative method is dynamic process simulation, which has also been recommended in the latest API guideline:1

“Conventional methods for calculating relief loads are generally conservative and can lead to overly sized relief and flare system designs. Dynamic simulation provides an alternative method to better define the relief load and improves the understanding of what happens during relief.”
This article gives an overview of the advantages of the approach, guidance on when and where to apply it, and describes a case study of a recent study at BP’s Lingen refinery in Germany.

Simplified methods for complex phenomena
Most events inside process units that lead to the opening of a relief valve are of a dynamic nature and lead to a relief load profile that shows a peak when the relief valves open and a more or less pronounced settling curve. This curve, as well as the pressure curve inside the process unit, is the result of many overlaid, non-linear phenomena inside the process unit. Any attempt to linearise these phenomena under-rates the dynamic aspects of the event. Conventional calculations use such a linearised approach and apply large safety factors to counteract the effects of simplification. This, in return, leads to the over-design mentioned in the API guideline cited above.

Dynamic process simulation
When applying dynamic simulation, it is possible to create a detailed model of the process unit and the dynamic phenomena that occur during the relief event. Therefore, it is possible to develop a realistic understanding of the overall relief behaviour of the process unit. The advantages of applying dynamic process simulation for relief load studies compared to the conventional mass and energy balance approach are:
• Better estimation of maximum flare load
• Better assessment of simultaneity of different peaks
• Analysis of effectiveness of planned measures, such as control and SV resizing.

Better estimation of maximum flare load
Applying dynamic simulation generally leads to a more realistic estimation of the maximum flare load from a process unit in an emergency situation; for instance, fire or power failure. This is due to the fact that dynamic simulation considers many phenomena that 
are not considered in the more conventional methods. Examples of these phenomena are:
• Thermodynamics of the system during each step of the scenario, and not only at the beginning and the end. Therefore, complex phenomena that have a significant effect on the results, such as evaporation of lighter components and its effect on physical properties, are considered
• Time-dependent effects of contri-butors to the pressure increase, such as valve closing times, control system action and limited steam availability
• Effect of cascade relief valves on the maximum flare flow.
As the API guideline states, in general, applying dynamic process simulation will result in lower estimates for the maximum flare load compared to conventional methods. Nevertheless, there could be situations where dynamic simulation could lead to a higher maximum peak flare load than the maximum average flare load calculated with conventional methods. This is an important aspect to consider when authorities base their operating permission for the plant on the maximum peak 
flare load.

Better assessment of simultaneity of different peaks
By estimating the flare load entering the flare system as a function of time, it is possible to predict potential simultaneity effects from different flare units. This infor-mation could be used to calculate cumulative flare load curves from different flare units. The application of simultaneity factors can therefore be omitted. Figure 1 shows the cumulative flare load curve from three process units. The maximum value of the cumulative curve is significantly lower than the sum of the maximum flare loads of each of the process units.

Analysis of effectiveness of 
planned measures
Measures to reduce the flare load in an emergency situation can be tested using the dynamic model of the process unit. Therefore, it is possible to test different safety concepts and to understand what is the maximum time allowed for control actions to take place.