Flare system design: Liquid pockets in flare headers

Flare and relief systems are the last line of defense against plant emergency overpressure and loss of containment. Therefore, the efficient design, management and maintenance of these systems are vital for safe operations throughout the lifecycle and operation of a plant.  

 

Flare header piping design is the key component of this system, and must consider any possible phase change, resulting in a multiphase flow regime in the flare header. Liquid carryover to the flare stack could compromise the overpressure protection, especially if condensation occurs after the knockout (KO) drum. Therefore, well-designed flare headers and integrity assurance are critical for the system’s overarching reliability. 

 

Pocketed sections in flare piping creates potential safety hazards, including accumulation, deposits and corrosive fluids that can hinder the flow needed during emergency demand, and potential slug formation, which can result in failure and loss of containment. 

 

Once the design proceeds into the detailed engineering project phase, the opportunity to modify or optimize the piping system design diminishes or triggers major change orders and delays to the project. This article focuses on areas where major benefits can be achieved early in the front-end engineering design (FEED) stage of the project. 

 

It has often been observed that flare headers in certain facilities contain these liquid pockets. This could be a result of restrictions on stack height or an attempt to curb costs associated with pipe support and structure. However, this practice contradicts best practices and guidelines given in international standards prohibiting the presence of liquid pockets in flare system piping. 

 

Flare header design: Typical orientation. Flare system piping orientation is a critical component of a plant, consequently leading to its development early in the engineering design phase. This helps avoid any potential surprises and major changes during the detailed engineering stage that can significantly impact cost and schedule.  

 

Within the flare system, the flare KO drum (FKOD) is used to separate liquid from vapor/gas. FKODs are essential in combating the risk of burning rain, as the flare stack is not designed to handle large volumes of liquid.  

 

Ideally, the main FKOD should be located closest (≤ 100 m) to the flare stack to ensure maximum liquid dropout. API 521 specifies that the slope requirement of 21 mm in 10 m (0.25 in. in 10 ft) is maintained for all laterals and headers. The slope should be towards the FKOD to avoid low points. Typical sloping requirements are shown in FIG. 1.  

FIG. 1. Typical flare header orientation. 

Adequate slope ensures that any condensate formed after passing through the FKOD will be effectively collected in the FKOD without creating any risk of liquid accumulation in the flare header. 

 

Flare header design: Noncompliant orientation. As noted above, the optimal positioning for the KO drum should be near the flare stack, which many older facilities overlook. This leads to difficulty in maintaining an adequate slope on the flare headers, consequently creating low points and/or liquid pockets. A diagram of a noncompliant flare header orientation is shown in FIG. 2.  

FIG. 2. Noncompliant flare header orientation. 

When water and heavy hydrocarbon components—if present in relief loads at relatively high relief temperatures—leave the main FKOD, they may condense, particularly in winter ambient conditions. As a result, liquid may pool inside the flare headers between the flare stack and the FKOD, particularly in the "liquid pockets" (low areas). Accumulated liquid has the potential to carry over to the stack and increase the back pressure on relief devices during relief. Therefore, a suitable liquid collection system design is required for such a design.  

 

Deviations from the typical design orientation as per API 521 can be attributed to a variety of design constraints. These constraints may include limitations related to high radiation, structural considerations and operational requirements. The following are some design constraints that may have influenced the design practice. 

• Flare radiation and dispersion: International and company standards provide specific radiation limitation guidelines that must be met. To limit the radiation at ground level, the stack height is adjusted. Nevertheless, high radiation levels at ground level may not necessarily be reduced due to limitations to stack height and/or relief flow. Limitations on stack height may also be imposed due to the proximity to sensitive government or aviation installations. In these circumstances, locating the FKOD farther away could be the only viable option. Flame-out dispersion analysis could also be a factor in this decision, as the presence of flammable or toxic clouds should not reach the FKOD. 

• Facility operations design preference: In some cases, the flare stacks are located far from the main processing facility (distances ranging from 1 km–2 km). The elevated flares, along with the ground flare (burn pit), are located in a discrete fenced flare area. In such circumstances, operations may have a preference to place the main FKOD within the battery limit of the processing plant for easy access and maintenance.  

• Cost avoidance/installation cost: If the flare area must be located farther away due to design constraints, the situation is made worse by running the flare headers on low-height sleeper racks (FIG. 2) to avoid the cost associated with the structural supports needed to hold the large-diameter pipes elevated to fulfill the sloping requirements. As a result, this design adds a low point in the main flare headers, thereby introducing unnecessary hazards of corrosion and liquid pockets in the flare header system.  

 

Such flare header design necessitates careful design of the liquid removal and recovery equipment, including a liquid accumulation vessel as well as pumps and necessary instrumentation. Experience has shown that these designs are inadequate and pose a serious hazard to the facility. Inadequate liquid hold-up volume and removal capacity could result in liquid carryover into the header or stack, resulting in slug flow. Slug flow may jeopardize the mechanical integrity of the system and cause burning rain. These designs have also been shown to exhibit inadequate slope due to sagging header pipes between supports, creating major corrosion issues.  

 

The drain vessel is connected to the main flare header through a small (4 in.–8 in.) diameter drainpipe. The efficiency of liquid separation and drainage depends on the fluid velocity and flow regime. Lower gas velocities tend to result in stratified or segmented flow regimes, thus helping liquids to drop out. Inversely, higher gas velocity entrains more liquid with vapor past the drain point towards the flare stack. Since the intention is to check the adequacy of existing collection systems, it is assumed that all condensed liquid will be drained into the collection system.  

 

Following a flaring event, the liquid flowrate and direction are primarily determined by gravity, the location and quantity of liquid holdup, condensation caused by ambient cooling, and the sloping of individual header sections in relation to the low point(s) and branch lines to the liquid collectors.  

 

Condensate formation depends on the type of relief stream, fluid velocity, flow regime, header length diameter, wind velocity and ambient conditions. Subsequent sections discuss the condensate volume calculation methodology and the appropriateness of the collection system. 

 

Condensate volume calculation methodology. Water and heavy hydrocarbon components present in the relieving streams drop out of the vapor phase due to ambient cooling, depending on the heat transfer rate through the header (insulation/heat tracing), ambient wind speed, and the temperature difference between relief fluid and air. It is recommended to check the condensation during and after an emergency relief event. 

              Condensation estimation during a relief scenario. Analysis was carried out using a flare network hydraulic simulation softwarea. The simulation was developed with the following considerations: 

1. The model was built using detailed composition basis (mole fractions), and appropriate enthalpy/VLE (Peng Robinson) and pressure-drop (Beggs & Brill) methods. 
2. Enable heat transfer with correct ambient conditions—i.e., lowest recorded ambient temperature, insulation, wind velocities.  
3. Consider all governing/applicable relief scenarios—i.e., high temperatures, rich hydrocarbon streams and/or with high water content (e.g., sour water stripper relief).  
4. Use site-specific meteorological and seismic data as per the company-issued design basis—run sensitivity cases [e.g., 3 ft/sec or average wind velocity as per site conditions (15 ft/sec and 32 ft/sec)]. 
5. Use the industry practice of a 20-min operator intervention response time to estimate the condensation quantity for a 20-min relief duration. 

 

                Condensation estimation at the end of a relief scenario. Since the driving force (differential pressure) is minimal at the end of the relief scenario, the relief stream stops flowing and the remaining vapor in the header gradually condenses out. To estimate the quantity of condensation, refer to the following procedure:  

1. Estimate the flare header pipe volume from the main FKOD to the stack.  
2. Identify the relief stream that condenses the most at a constant header backpressure. Check for steams with high water content and run a sensitivity analysis.  
3. Calculate the liquid density at constant header backpressure using a selected relief stream composition. 
4. Use liquid and vapor density and header empty volume to estimate the volume of condensate (Eq. 1): 

    

Liquid collection system adequacy verification. Based on condensate volume calculations, a vessel could be sized for a holdup time of 20 min–30 min, an average time for operator intervention. The collection vessel must have an automated liquid removal mechanism (e.g., a pump or control valve) to transfer the liquid to an appropriate destination (i.e., to the FKOD, slops, closed drain or burn pit). It is important to have a low- and high-level indication on the plant distributed control system (DCS) and appropriate instrumentation to automate the operation of liquid removal. The low-point liquid collection vessel/equipment is detailed in FIG. 3.  

FIG. 3. Low-point liquid collection vessel/equipment.    

The absence of automated instrumentation for the drainage system poses a significant risk. Manual operation, by its very nature, is prone to human error, particularly during emergency situations. Any malfunction or delay in the automatic liquid removal operation could lead to liquid carryover to the flare stack, potentially causing safety hazards and operational disruptions. 

 

Redundancy in the liquid removal mechanism—i.e., motor-operated pump or control valve (on/off)—is required. Two pumps—one pump with a secondary dump valve—or two valves could be used, depending on the system hydraulics, design and space availability.   

 

A case study was undertaken on an existing facility to assess the adequacy of the existing collection system in terms of capacity and automation. Additionally, the study aimed to offer initial guidance to design engineers on how to develop an appropriate collection system. 

 

CASE STUDY 

 

A low-pressure (LP) flare header was assessed at one of the existing sites with a header orientation like the one shown in FIG. 2. The analysis was carried out following steps provided in the “Condensate volume calculation methodology” section using simulation softwarea, which shows the condensate formation in the header downstream of the KO drum. The LP flare header system’s specifications are detailed in TABLE 1.  

The length of the header from the FKOD to the flare stack was considerable, thereby increasing the likelihood of condensate formation in this header. TABLE 2 includes only the relief scenarios in which condensation is formed. Since the phenomenon occurs downstream of the KO drum, an adequate collection system would be required to remove the condensate from the header. Two cases were considered to estimate the condensate formation (i.e., during and at the end of the relief scenario) for relevant overpressure relief scenarios. 

 

There was no insulation or heat tracing on the flare header. Analyses were done for various wind velocities (high, medium and low) to estimate the condensate formation. More condensate was formed at a higher wind velocity; however, it was separated out at the main FKOD, resulting in less condensation downstream of the FKOD. Medium wind velocity yielded higher condensate after the FKOD, whereas, due to reduced heat transfer at a low wind velocity, minimal condensate was formed. 

 

Case A: An estimation of condensation during relief. TABLES 2–7 provide process parameters for the applicable overpressure scenarios downstream of the FKOD. The scenarios were simulated at various ambient conditions provided by site meteorological data for consistency. Hydraulic simulations were run using simulation softwarea. 

TABLES 2–7 represent vapor fraction at different wind velocities and amounts of condensate, with the assumption that the scenario may last for 20 min.  

 

All applicable overpressure scenarios were screened from the facility flare and relief design documentation to ensure that all possible outcomes were considered. 

Case B: Estimation of condensation at the end of relief. TABLE 8 provides an estimation of condensation at the end of the relief scenario. The total power failure overpressure scenario yielded maximum condensate; therefore, it was selected for this case. To calculate the volume of liquid formation within the header, Eq. 1 was used.  

The end of relief scenario’s condensation volume was estimated to be 20% of the condensation during the relief scenario. In most cases, condensation formed during the relief scenario formed the basis for the design of the collection system. However, it is recommended to consider both cases to determine the sizing case for the liquid collection system. TABLE 9 shows the condensate volume for Cases A and B. The sizing for the existing collection system is detailed in TABLE 10.  

The existing collection system is inadequately designed and lacks the necessary automation for liquid drainage, posing a significant risk. It was recommended to increase the collection system size to accommodate the formed condensate and include level alarms and centrifugal pumps operating in lead-lag automatic operation, as shown in FIG. 3. 

 

Takeaway. It is imperative to ensure that the flare header design is optimized during the early stages of the engineering design. The flare and relief system is the last line of defense against overpressure emergency situations; therefore, a robust design is critical for the safe operation of the facility. Care must be exercised to minimize any design deficiency that would jeopardize the safety of equipment and personnel.   

 

The main FKOD should be in close proximity to the flare stack, considering personnel access for maintenance and thermal radiation from the flare. The KO drum must be sized to separate liquids from the gas stream and to hold the maximum amount of liquid during a relief scenario. 

 

The flare header design with liquid pockets (FIG. 2) must be avoided. The design may save on capital expenditure; however, experience has shown that this design is prone to developing corrosion, requiring frequent inspection and repair and posing the risk of liquid carryover to the flare stack.  

 

Operating facilities with such a design typically exhibit a lack of proper slope, creating liquid pockets that are subject to the accumulation of deposits and corrosive fluids that might lead to high under-deposit corrosion risk. The piping systems should be inspected according to the company’s onstream inspection dead-legs inspection program. Nondestructive testing may have to be conducted to identify susceptible locations and select the corrosion monitoring location for longer piping circuits. The identified locations should be inspected using quantitative inspection methods, and header sections identified with a low thickness may need to be repaired/replaced. 

 

Disclaimer. The information contained in this article represents the current view of the authors at the time of publication. Process safety management is complex, and this document cannot embody all possible scenarios or solutions related to compliance. This document contains examples for illustration and is for informational purposes only. Saudi Aramco and Siemens Energy make no warranties, express or implied, in this article. Furthermore, any adverse impact resulting from the implementation of the solutions provided in this document may not be attributed to the authors or their respective companies. 

 

NOTE  

^a AspenTech’s Aspen Flare System Analyzer 

 

ABOUT THE AUTHORS  

Hasan Ali Amin earned an MS degree in chemical and petroleum engineering from the University of Calgary, Canada. He joined Saudi Aramco in 2015 and works in the Flare and Relief Systems Group as the Lead Engineer. Amin has more than 25 yr of process engineering experience in the oil, gas and petrochemical industries with leading engineering, procurement and construction (EPC) companies in Alberta, Canada. 

  

Mahdi AlDajani is a Supervisor Engineer in the Flare and Relief System Group at Saudi Aramco. He earned a BE degree chemical engineering from King Fahd University of Petroleum and Minerals (KFUPM), Saudi Arabia. He has more than 15 yr of experience working within the area of process safety and risk assessment, and has extensive experience in pressure relief and flare design for existing facilities and greenfield projects.  

 

Praveen Dhote is a Process Safety Team Leader at Siemens Energy. Dhote earned a BTech degree in chemical engineering, and has 23 yr of experience specializing in process safety management. He graduated from Laxminarayan Innovation Technological (LIT) University in Nagpur, India. He has extensive experience in pressure relief and flare system design and revalidation. Dhote is a certified functional safety engineer, chairs HAZOP and SIL workshops, and is a certified Hydrogen Safety Engineer from Ulster University, Ireland. 

 

Related Articles