Flare network low-temperature analysis

Accurate material of construction (MOC) decision support

A key challenge in flare network design is determining the minimum metal temperature seen at certain at-risk points in the system.

If the temperature is too low for the material of construction (MOC) there is a risk of brittle fracture. On the other hand, unnecessary use of low-temperature metals can add significant capital cost.

PSE's gFLARE technology is specifically designed to provide accurate metal temperature modelling based on rigorous thermodynamics. Read our case studies to see the impact of high-fidelity modelling on real projects.

Example

The flare network design for a gas processing plant shown below, designed and sized at FEED stage, is now undergoing detail design.

Flare network

It is suspected that metal temperatures in the flagged section of line will be below safe limits. The key questions to be answered are:

  • what is the minimum temperature seen here?
  • for how long does it last?
  • what length of this pipework needs to be made of low temperature grade steel

Existing design

During the FEED, it was determined that the main header and KO drum should be constructed from Low Temperature Carbon Steel rated to −46°C.

However the simultaneous depressurisation of two blow down segments results in gas entering the flare network at an entry temperature of -71°C.

Subsequent analysis with the standard steady-state flare network design tool shows that metal temperatures in flare network header pipe and KO drum are significantly below the minimum −46°C design temperature.

Approach

The flare network configuration was imported into PSE's gFLARE, selecting only the active branches, which resulted in the flowsheet shown on the right.

A comparison between the base case and the steady-state tool results for selected elements shows good agreement:

Pressure [bar abs]gFLARESteady-state tool
Flare stack (inlet) 1.192 1.192
Pipe segment 12 (inlet) 1.488 1.464
Pipe segment 11 (inlet) 1.516 1.511
BDV 1 - Back pressure 1.719 1.716
BDV 2 -  Back pressure 1.753 1.749

Analysis of the steady-state results shows that the temperatures are indeed below the LT Carbon Steel embrittlement temperature of −46°C, reaching a minimum of −69°C in the pipe entering the KO drum and −66°C near the inlet of the drum itself.

Steady-state vs dynamic simulation

Steady-state analysis does not provide a reliable prediction of metal temperatures because the relief flow is assumed to continue at the peak flow forever. In reality, relief flows rise to a peak for a short period then taper off.

Steady-state analysis thus ignores important thermal inertia effects, and does not accurately account for heat flow in from the surroundings. Dynamic simulation is required for accurate analysis.

Results

The same flowsheet was executed in dynamic mode to show the extent and duration of the constraint violation. The results are shown in the plot below:

Flare network dynamic temperature prediction for concurrent blowdown

Initially, the thermal inertia of the pipe metalwork reduces the cooling effect. The metal temperature in Pipe segment 11 drops to a minimum of −54°C and then rises as the relief flowrate drops and heat comes in from the outside.

However despite the reduction in the minimum temperature, it still violates the limit. It is now worth examining the effect of staggering the blowdown flows so that the second starts after the flow from the first has begun to subside.

The results from the staggered case can be seen below:

Flare network dynamic temperature prediction for staggered blowdown

In this case, the initial temperature reduction in Pipe 11 is slower because of the lower flowrate. By the time the second blowdown occurs, the heat flow into the system compensates for the cooling of the blowdown flow; theLT Carbon Steel embrittlement temperature of −46°C is never reached.

Conclusions

Until now much safety design and analysis has been done using steady-state or "pseudo-dynamic" trial-and-error modelling that fails to capture the true behaviour of the system.

It can be seen here that by the simple expedient of staggering the blowdowns times the system can be made to operate within safe limits.

This type of behaviour cannot be predicted using steady-state methods. indeed, as most safety work deals with transient conditions, it is essential to use full dynamic models with proper modelling of pressure-flow behaviour.

Apply high-fidelity dynamic modelling to your MOC decisions

The use of a rigorous model-based safety approach based on our gFLARE technology and coupled process and flare network models can result not only in improved process safety but also an improvement in operations and a considerable reduction in CAPEX.

Our website uses cookies so that we can provide a better browsing experience. Continue to use the site as normal if you're happy with this or find out more about cookies

OK