Process Systems Enterprise Limited
email this page print this page
pdf overview

Model-based Safety

The study enabled us to save $1.5m on the cost of this vessel.

— Stephen Leng, BP Operational Excellence

 

Case: knock-out (KO) drum materials of construction selection

In constructing the relief system for the Angola Block 18 offshore development, BP urgently required high-accuracy information to support a decision on material of construction for the system knock-out vessel.

They applied PSE's flare system models to determine the extent of cold penetration into vessel and surrounding pipework walls. The diagram on the right below (click for detail) shows the areas modelled.

System detail

Flare stack with depressurization

The challenge: feed temperatures of below –100°C

Model formulation

Depressurization 3-D wall discretization and heat fluxes for temperature modelling
3-D discretisation and heat fluxes – click for more detail

Depressurization inconel vs. carbon steel decision
Decision: Inconel vs. carbon steel?

 

Scenario and results

Depressurization worst-case scenario
Worst-case scenario – two successive blowdown events

Flare stack with depressurization
Tube wall temperatures at inlet and outlet for worst case

Flare stack with depressurization
Azimuthal temperature profile over time

Depressurization - temperature color map through knock-out vessel wall thickness
Temperature color map through knock-out vessel wall thickness

The vessel is 15m in length, and has to handle incoming gas at around -100°C. If it could be shown that temperatures in the vessel wall remain above 230 K at all times during depressurisation, the vessel could be constructed from carbon steel with an Inconel lining.

However steel may become brittle and fail at very low temperatures. Thus if the temperature approached or exceeded the limit, it would be necessary to construct the vessel out of the expensive Inconel alloy.

The tank usually contains some residual liquid, which is instrumental in reducing extremes of temperature.

Modelling approach

The model used the pre-production vessel model from PSE's gFLARE library. This included:

  • 3-dimensional (axial, radial and azimuthal) modelling of heat conduction through the steel vessel walls and Inconel liner.
  • 3-dimensional modelling of the residual liquid in the tank.
  • 1-D energy model for gas phase, allowing for axial variation.
  • Predictive heat transfer correlations for:
    • solid-to-ambient heat transfer coefficient
    • gas-to-liquid/solid heat transfer coefficient

The model was executed for various scenarios, including different amounts of residual liquid and two blowdown events in sequence.

Results

The results of the simulation are an accurate steel wall temperature map that shows temperatures at any point in the vessel walls over the elapsed time of the depressurisation. Examples are shown on the right.

The results showed that, despite the extreme temperature of the gas entering the vessel, at nowhere does the steel wall temperature ever approach the 230 K limit for carbon steel.

Conclusions

The study verified lower-cost design, showing conclusively that it was unnecessary for the vessel to be constructed from Inconel, and the less expensive carbon steel option could be used with an Inconel lining.

This saved $1.5m on the vessel construction, and allowed the vessel to be constructed and delivered much earlier than it would otherwise have been.

Modelling timescale was of the order of 15 days elapsed, well within the purchasing decision timeline.

Following this study, the same procedure was also applied to the upstream pipework.

With thanks to BP