Complex Flare Network Analysis Case Study

The BP Texas City incident in March 2005 demonstrated the catastrophic consequences that can result from an improperly designed effluent handling system. Recommendations made by the U.S. Chemical Safety and Hazard Investigation Board following their incident investigation prompted many petrochemical companies to engage in reviews of their effluent handling and flare systems.

In this newsletter, Georges Melhem , Ph.D., FAIChE, and Neil Prophet present their approach for performing a comprehensive flare systems evaluation for a large oil refinery whose flare network had grown too complex for their in-house software to model effectively. This challenge prompted the refinery to explore more advanced external solutions.

The Challenge

Our client had a very complex flare network which had been modified numerous times over the lifetime of the facility. It had six separate flares, three main headers, and hundreds of relief devices that discharged into the system. As the flare system was modified over the years, multiple cross-connection points were added between the headers in an effort to balance the flow rates through the headers with minimal piping changes.

The problem was that the system had become so complex that the tools our client was using to evaluate the flows through the flare network could not adequately model the system. Management no longer had confidence that the model results reflected the actual network performance and therefore, could not be sure the system would perform properly in the event of a global relief scenario at the facility.

Our Approach

The ioMosaic team developed an accurate model of the flare network and evaluated the performance during a total refinery power failure scenario.

To do this, they constructed a model of the flare network in Process Safety Office? SuperChems?. The backbone of the model included developing piping isometrics for all of the relief devices, headers, and flares involved in a total refinery power outage based on the actual field piping isometrics provided by the client. All of the relief devices associated with this scenario were also modeled using the valve details provided by the client.

Once the physical details of the flare network had been entered into the program, the network was divided into sections to assist with the complex calculations. Points where the flows could cross between headers were identified as Nodes. Some of these Nodes had relief devices providing flow into the Node and others were only potential cross-over points between the different headers, but no new flow was added. For each Node where relief devices provide flow into the Node, the relief devices were grouped together. The network ended up with 11 groups of relief devices and 20 network Nodes. The Groups of relief devices ranged from just one relief device to 19 devices in each group.

A back-pressure curve was generated for each relief device using the specific characteristics of the relief valve and the chemical components flowing through it. Next a back-pressure curve was generated for each Group of relief devices. This provided an initial look at how the relief devices would perform when multiple devices were flowing simultaneously.

All of the network Nodes were then defined with interconnecting piping isometrics and information about where flow could come from and go to for each node. In many instances, flow could go either direction through a crossover pipe connecting two headers depending on what the pressure was at each of the connecting Nodes. Other software packages ran into trouble without a defined flow path through each of these connection points. However, it wasn’t possible to confirm the flow direction until the pressures were identified.

Once all of the Nodes were defined, Process Safety Office? SuperChems? generated the equivalent of backpressure curves at multiple temperatures and pressures for each Node, called Flow Maps. These Flow Maps were used to converge the network solution. The software provided temperatures and pressures for each Node and the flows through the network, including which direction the flow was going through each of the network cross-over connections.

The final step in the analysis was to review each of the groups of relief devices and determine how much flow could pass through each device with the back-pressure at the first network Node as determined above. These relief device flow rates could then be compared to the required flow rates for this particular relief scenario to determine if the relief device and network piping could meet the demand.

Once the model was developed, modifications to the network could be easily examined. If safeguards were present that eliminate the relief load from a particular device, that device could be disabled by a single click. Changes to the type of relief device (conventional, bellows, or pilot) were easily examined. The hydraulic analysis output identifies the pressure drops through each segment of piping, allowing the user to quickly determine which sections of piping should be modified to achieve the greatest performance improvement within the network.

Based on the changes made, it was possible to re-evaluate the flows through the network. In some cases, if flows were removed from certain groups, then the pressure would drop, and flow might reverse direction through one of the cross-over connections. Process Safety Office? SuperChems? software identified these changes in flow direction and provided a new solution to the flow network.

In Conclusion

The ioMosaic team was able to produce a dynamic model of a very complex flare network where all other software packages came up short. Once developed, a model is easily modified to stay current with changes to the system.

To learn more about Process Safety Office? SuperChems?, visit our website > https://bit.ly/2X4mqF8

Speakers Showcase - Practical Guidelines for PRV Stability

Watch this 46-minute presentation by Georges Melhem , Ph.D., FAIChE, to learn practical methods for the identification and evaluation of high risk relief systems installations. He will demonstrate advanced developments in PRV stability and include working examples on evaluating and performing engineering analysis when PRV instability is suspected using Process Safety Office? SuperChems? software.

Your Facility Can't Afford to Overlook Past Safety Lessons Article

Understanding regulations is just the beginning; the true test is how your team effectively integrates lessons learned into their daily operations. A common critique of Process Safety Management (PSM) training is its tendency to prioritize regulatory content over practical, real-world applications. Read this article to learn the impact on safety when we fail to apply past lessons learned.

To read the article, visit our website > https://bit.ly/4hppK4d

How We Can Help You

Emergency Relief Effluent Handling System Design

Our team has decades of experience performing PRFS analysis and design and methods to maximize existing flare structures.

Pressure Relief and Flare System Design

Our risk-based approach helps mitigate near-unventable scenarios to a tolerable level of risk and develop economical designs for more credible events.

Process Simulation

Better and more accurately evaluate hazards in your oil, chemical, pharmaceutical, or LNG facility with an accurate process simulation.

Relief Header and Flare Analysis Systems

Delivering properly designed pressure relief systems for refineries and chemical plants that save both money and time.

To learn more, visit our website > https://bit.ly/3eKzTJA

#ReliefDesign #FlareNetwork #RiskManagement

Copyright ? 2025 ioMosaic Corporation. All rights reserved