Format And Type Fonts CHEMICAL ENGINEERING TRANSACTIONS VOL. 39, 2014 A publication of The Italian Association of Chemical Engineering www.aidic.it/cet Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Peng Yen Liew, Jun Yow Yong Copyright © 2014, AIDIC Servizi S.r.l., ISBN 978-88-95608-30-3; ISSN 2283-9216 DOI: 10.3303/CET1439266 Please cite this article as: Deilamani K.S., Rajabi M., 2014, Black smoke assessment of low pressure flare in South Pars gas field off-shore platform based on FLARENET simulation and flied experiences, Chemical Engineering Transactions, 39, 1591-1596 DOI:10.3303/CET1439266 1591 Black Smoke Assessment of Low Pressure Flare in South Pars Gas Field Off-Shore Platform Based on FLARENET Simulation and Flied Experiences Kaveh Soleimani Deilamani* a , Mehdi Rajabi b a Process Design and Flow Assurance Engineer, Iran International General Contractor (I.G.C), Tehran, Iran b Production Manager, South Pars Gas Field Phase 1 Platform, Pars Oil and Gas Company (POGC), Assaluyeh, Iran k.soleimani@igc-ir.com Flaring is the combustion process used for the safe disposal of large quantities of waste gases and vapours in the petroleum and natural gas industries. Due to the fact that toxic and greenhouse gases are the main products of flaring which has devastating impact on environment and also ruin distinction of the operators of the plant specifically when accompany with black smoke, in the recent years, researchers and policy makers are focused on these issues by finding solutions and legislating protocols and environmental regulations. The present study revolved around potentials of smoking of Low Pressure (LP) flare system of South-Pars Phase 1 platform to throw some light on the exact operational and theoretical reasons and ultimately propose a practical solution in order to sort out the issue considering off-shore restrictions. In addition to investigating possibility of liquid carry over with waste gas to flare tip, the LP flare System is simulated for continuous scenario for evaluating composition of waste gas, moreover, the adequacy of knock out drum is assessed regarding its size and efficiency of separation of gas and liquid. The study showed that the low momentum of waste gas mixture is the main reason of smoking and injection of air as assistance creates effective mixing and smokeless flaring. 1. Introduction Generally, Flaring is the inherent part of every oil and gas plants to supply safety for process and prevent fatality and damage to the asset. However, it brings about serious environmental issues if it is not design properly. It is estimated that maximum 200 billion cubic meter gas is being flared every year. Incomplete burning would always accompany with noticeable black smoke which could be tarnished the reputation of the plant or platform in addition to association with environmental problems due to releasing toxic gases. Therefore, necessary measurements need to be taken on one hand, by officials by means of speculating strict regulations and one the other hand, by researchers to propose technical solutions to eradicate this issue. Brzustowski (1976) stated that the most common practical method of suppressing smoke in flares is the addition of steam. The steam has the greatest effect when it is injected into the flame at/or just above the discharge plane of the flare tip, in such a way that it can also entrain some air into this region. When steam is not available compressed air could be used. Shore (1996) presented that some gases are more likely to be smoky because they contain a greater proportion of carbon than others or because their cracking temperature is lower than others. Smokeless flaring of low-pressure gas is significantly more difficult than smokeless flaring of high-pressure gas. When flaring a high pressure gas, not only is the gas lighter, but the gas velocity at the flare tip usually creates sufficient turbulence to ensure the proper air and gas mixing required for smokeless combustion. While, the flash gas coming off a low-pressure separator or any other equipment would have higher molecular weight requiring more air for complete combustion. If the low-pressure gas is allowed to burn without the introduction of sufficient outside air or steam, the flare tends to smoke excessively. In addition, the flame becomes too close to the flare tip causing flare tip damage (Obawole et al., 2001). Mashour (2009) has patented a new smokeless flaring technology using high-pressure air. This technology is called High Pressure Air Assist System (HPAAS) and has been 1592 successfully implemented on dozens of flare systems in Saudi Arabia. It uses supersonic air injection nozzles to inspirit smokeless air at a much higher efficiency than any previous smokeless high-pressure air-assisted flare technology. The main objective of the present work is to investigate various reasons for black smoke formation of flare on South Pars Phase1 platform and find the solution to sort out the issue applying field experiences and results of flare network simulation with Aspen Flare System Analyser V7.1. 1.1 Process Description The South Pars Gas Field has been initiated for the exploitation of a 1,000 MMSCFD of gas and condensate reservoir located in the Persian Gulf. The phase 1 facilities are designed to allow free flow of well fluids from the wellheads to the refinery without any requirement for compression or pumping. The fluids from the wellhead platform are routed to a manifold which enables production from either platform to be routed to each processing train. Each train consists of: Primary separation/cooling, gas dehydration and condensate dewatering in which gas is processed to reach certain specification in order to transport into on-shore facilities. Dehydration is accomplished by contacting the gas with lean Triethylene Glycol (TEG). The water rich TEG is sent to the Glycol Regeneration Package (GRU) where the water is stripped from the TEG so it can be reused. Figure 1 indicates the schematic diagram of Phase1 main process and GRU facilities. 1.2 Phase 1 LP Flare System The LP Flare allows the safe flaring of hydrocarbon gas from equipment with a design pressure of less than 16 barg. A continuous flow of low pressure gas is come from the produced water degassing vessel and the GRU. Also, the LP Flare is continuously purged with fuel gas. The LP flare system consist of flare headers which collects hydrocarbon releases from LP process equipment and routes them to LP Flare KO drum. The LP flare KO drum is designed to separate entrained liquid droplets greater than 600 microns in diameter from the flared gas for all flaring conditions, with maximum liquid level in the drum. 2. LP flare black smoke Noticeable black smoke has been observed during continuous flaring which is very harmful for the environment. Figure 2 illustrates the black smoke in the flaring. In order to solve the issue, a comprehensive study has been carried out by means of brain-storming in order to find out the probable reasons. Based on the investigations main reasons of black smoke are categorized as following; 2.1 Liquid in the Vent Stream This is very ubiquitous to consider presence of liquid in vent stream when smoking is observed in the flaring. Liquid in the vent stream can extinguish the flame or cause irregular combustion and smoking. According to the site visit, when LP flare is unloaded completely, at the first hours and when the liquid level in KO drum is low enough, no black smoke is observed, however, as soon as liquid level in the LP KO drum increases, combustion is accompanied by black smoke. The possible reasons for presence of liquid in the LP KO drum is low enough, no black smoke is observed, however, as soon as liquid level in the LP KO drum increases, combustion is accompanied by black smoke. The possible reasons for presence of liquid in the LP Figure 1: Process Block Diagram of main process facilities and GRU of phase 1 platform 1593 Figure 2: Black smoke in flare of phase 1 platform flare system could be inadequate size of KO drum and liquid levels in the drum which could reduce the volume available for vapour/liquid disengagement. This should be considered when defining liquid level high (LAH) for the KO drum. As a matter of the fact, when LAH which causes pump to start is not designated properly, the possibility of liquid carry over with gas is escalated. 2.2 TEG presence in vent stream Because of the sensitivity of TEG based on the operator experience, it is specifically considered as one separated item. In GRU package, TEG enters still column containing two parts of random packing. It also passes re-boiler working at 204 °C prior to stripping column. The stripping gas (LP fuel gas) is also enter the surge vessel continuously and flows up through the structured packing in stripped column in counter- current flow with glycol. During this mass transfer process most of the left over water in glycol will be removed. The stripping gas is routed to LP flare header at the outlet of still column. Glycol could also be found on LP flare network from another source in GRU package. This is the glycol flash vessel in which its pressure should be maintained at 5.2 bar with LP fuel gas by means of a pressure control valve. 2.3 Inadequate Mixing and incomplete combustion When hydrocarbon composition is the main material sent to flare tip, one of the factors that control black smoke is the H/C weight proportion. The more the proportion value, the less black smoke is formed. Regarding gas mixture contains different hydrocarbons with H/C smaller than 0.25, smoke would be generated during burning. The main reason for this is inhomogeneous mixing of waste gas. According to the API 521, one of the important criteria for non-smoke flaring process is the heating value of the mixture. Gases that have a high enough heating value which is usually greater than 200 Btu/Scf (7,443 kJ/m 3 ) for unassisted flares sustains combustion on their own without any auxiliary fuel additions. 2.4 Waste gas momentum It has been studied that the momentum of waste gas playing significant role in burning operation. If the waste gas pressure (momentum) is not adequate to provide smokeless burning, other energy sources (e.g. steam or air), or a combination of energy sources can be used to compensate low momentum of gas and make the burning complete. 3. Flare network simulation In order to find out the real composition and also the heating value of the burning mixture on the downstream of KO drum, the LP flare system is simulated by means of Aspen Flare System Analyser V7.1. Phase 1 LP flare system comprises of two headers called continuous and non-continuous which are connected to LP KO drum separately. The adequacy of LP flare system is checked with only continuous sources to the flare header, since in non-continuous mode (emergency situation), smoking is not a concern. The specifications and operating conditions of continuous sources tabulated in the Table 1. LP Flare network is simulated applying mentioned data in the Tables 1 and available compositions. Also as- built piping sizes, elevation, material and thicknesses are utilized in simulation. The LP flare is sized to give low back pressures in the header equal to 0.013 barg. Figure 3 shows flare network simulation in Aspen Flare System Analyser V.7.1. 4. Results and discussion 4.1 Simulation Results The LP flare network is simulated to find the composition of fluid at flare tip and ultimately evaluate the heating value of the mixture. The data of each source are extracted from project documents and a 1594 scenario is defined in the software that includes all of the continuous sources. The pressure of the fluid at the outlet piping from Table 1: Specifications of continuous sources to LP flare network LP Continuous Sources Pressure (barg) Temperature (°C) Mass Flow (kg/h) Composition Purging fuel gas 3 20 0.7 Fuel gas composition Produced Water Package Train 1 & 2 2.8-3.2 20 10 Fuel gas composition GRU-Still Column-Train 1 &2 0.15 85 1,280 Stripping gas composition at outlet of still column GRU-PCV-Train 1&2 1.01 80.8 160.5 Vapour composition of flash vessel LP KO drum to flare tip is low enough that the gas mixture could be considered as ideal. The net lower heating value of the burning mixture is calculated 16,560 kJ/m 3 which indicates that the heating value of the burning mixture generated from LP flare network is more than criteria (7,443 kJ/m 3 ) and doesn’t need extra injection of gas with higher heating value. 4.2 KO Drum Sizing To ensure whether liquid droplets carry over with gas in the KO drum, the sizing of existing equipment is checked with the common procedure elaborated in the API standard 521. It is assumed that the relief time to be 20 minutes and liquid that accompany with vapour release has volume of 1.5 m 3 . Also, the Maximum relief load during emergency situation is 4,802 kg/h. the vapour density and viscosity is 0.74 kg/m 3 and 0.012 cp respectively. Liquid flow rate is considered to be 10 kg/h with density of 985.6 kg/m 3 . The Drop out velocity is then evaluated from Eq(1). (1) Where, Uc is the drop out velocity, dp is the particle diameter and C is the drag coefficient. The drop out velocity is found 3.41 m/s. Sizing of knock out drum is continued applying trial and error procedure regarding diameter and length of the drum. Here, the length and diameter of existing drum is used to check whether its size is adequate for the worst scenario occurring during emergency situation (D=1.1 m and L=3.3 m). Liquid volume in the KO drum would be liquid streams that accompany a vapour release and condensate that separates during a vapour release; AL1 and AL2 signify these parameters and calculated as below; (2) (3) (4) The vertical space for vapour flow (hv) is found 0.583 m. Liquid drop out time is then found utilising vapour space and drop out velocity equal to 0.17 s. The vapour velocity is also calculated as following; (5) Where Rv is the vapour volumetric flow rate. The key factor in sizing the knock out drum is that the minimum time for vapour passing the drum from inlet nozzle to outlet nozzle (called T/T length) should be equal to the drop out time of liquid droplets accompany with vapour, the minimum length of the drum is then found; (6) 1595 It should be mentioned that Lmin must be less than or equal to the assumed cylindrical drum length, L; otherwise, the calculation must be repeated with a newly assumed cylindrical drum length. The minimum required length for the drum in this study is smaller than assumed (existing) length, therefore, the existing KO drum size is adequate to handle relief stream during emergency and continuous circumstances. Figure 3: Part of LP flare network simulation area in FLARENET software 4.3 Modification in GRU package During a renovation of platform in a frame of revamping project, several modifications have been carried out in the GRU package and many of the equipment are replaced with new ones. Moreover, the adequacy of mentioned equipment regarding size, internal and etc. are checked and improved in the case of inappropriate situation. As an example, the still column packing were replaced with structured packing Mellapak 250Y which alleviate the flooding of the bottom packed section and make contribution for effective TEG and gas separation. A new pressure control valve is also provided for fuel gas routes to surge vessel and still column. Even after revamping project, the black smoke is observed in the LP flare and the presence of TEG on the downstream of KO drum is not proven by the field experience. 4.4 Gas Momentum The continuous operating conditions for Phase 1 LP flare is often low flow rates composed of purge gas (LP Fuel Gas), control valve leakage, or tank vapours. This mixture of gas exits at a low velocity and smokes continuously. A common smokeless method for low-pressure flares is to use a low-pressure air- assist. Despite high heating value, due to low pressure in the LP flare network of phase 1 platform the momentum of waste gas is not adequate and it burns without introduction of outside air or steam, thus, it needs assist media. Because of the off-shore restriction, using steam is not possible on phase 1 platform, thus, air assisted is more preferable. Air-assist flares use blowers to force air to the tip which is designed to promote air-fuel mixing and provide stable burning. The supplied air adds momentum and serves as a portion of the required combustion air. 5. Conclusion Recent changes in the attitudes and policies of individuals and agencies worldwide have resulted in the need for fewer smoking flares, e.g. one of the goals of the Kyoto Protocol is a reduction in worldwide flaring. A good flaring management could lead to reduce the emission of toxic gas which ultimately minimizes the impact of flaring on the environment. One of the basic principles to limit the hazardous gas emission in the case that no flare gas recovery projects are feasible is to make flaring combustion as efficient as possible. Based on the present study, Incomplete burning on the flare tip, presence of liquid in the flaring gas, low heating value of the burning product and low gas momentum are the most common reasons for observing black smoke during flaring. The mentioned reasons are investigated for LP flare system of phase 1 platform specifically with in-depth knowledge including both theoretical and field experiences of veteran operators and it is concluded that the most probable reason is the low momentum of flaring gas mixture. According to common practice when waste gas pressure is low, steam or air need to 1596 be provided as assistance to improve the level of momentum and produce complete combustion and smokeless operation. Injection of steam is not feasible on phase 1 platform due to some reasons including unavailability of steam on the platform, steam boiler and steam delivery system specifically for the flare system is not cost effective and lack of space for required equipment and facilities. Therefore, the most suitable and cost effective air assisted design considering existing facilities and equipment with minimum changes and modifications is applicable for South Pars phase 1 platform. References Amyotte P., Rigas F., 2013, Applications of Process Safety Concepts to the Hydrogen Economy, Chemical Engineering Transactions, 31, 31-36 . API Recommended Practice 520, 2000, Sizing, Selection and Installation of Pressure Relieving Devices in Refineries, Seventh Edition, American Petroleum Institute, Washington, DC, USA. API Recommended Practice 521, 2007, Pressure Relieving and Depressurizing System, Fifth Edition, American Petroleum Institute, Washington, DC, USA. Bader A., Baukal C.E., Bussman W., 2011, Selecting the Proper Flare Systems, Chemical Engineering Progress, July, 45-50. Brzustowski T.A., 1976, Flaring in the energy industry, Prog. Energy Combust. Sci., 2, 129-141. Green D.W., Perry R.H., 2008, Perry’s Chemical Engineers Handbook, 8 th ed., McGraw-Hill, New York, USA. Hong J., Baukal C.E., Schwartz R., Fleifil M., 2006, Industrial-scale Flare Testing, Chemical Engineering Progress, May, 35-39. Hong J., Baukal C.E., Bastianen M., Leary K., 2007, New steam assisted flare technology, Hydrocarbon Engineering. Klooster H.J., Vogt G.A., Braun G.F., 1975, Optimizing the design of relief and flare systems, Chem. Engg, Prog.71, 39-44. KLM Technology Group, Flare selection and sizing (Engineering Design Guideline), Practical Engineering Guidelines for Processing Plant Solutions Mashour M., Smith S., Palfreeman N., Seefeldt G., 2009, New Technology: Saudi Aramco High Pressure Air Assist System (HPAAS) for Upgrading Existing Flares to Smokeless Combustion, International Flame Research Foundation 16 th Members Conference. Combustion and Sustainability: New Technologies, New Fuels, New Challenges, USA. Mashour M., Smith S., Palfreeman N., Seefeldt G., 2010, Success Stories: Saudi Aramco High Pressure Air Assist System (HPAAS) for Smokeless Flaring, American Flame Research Committee Pacific Rim Combustion Symposium, 26-29 September 2010, Maui, Hawaii, USA. Obawole A.B., Onwuka O.C., Jan R.J., Martin M.J., McSorley K.P., 2001, Key design parameters for smokeless flaring of Low pressure gas, SPE/EPA/DOE Exploration and Production Environmental Conference, 26-28 February 2001, San Antonio, Texas, USA. Salooja K.C., 1974, Improved control of smoke, combustion efficiency and flame stability, J. Inst. Fuel, XLVII, 203 Shore D., 1996, Making the flare safe, J. Loss Prey, Process Ind., 9(6), 363-381. Sinaki S.Y., Tarighaleslami A., Jafarigol F., 2011, Study of external costs of flare gases using Asian development bank method, Chemical Engineering Transactions, 25, 39-44 Swithenbank J., 1972, Ecological aspects of combustion devices (with reference to hydrocarbon flaring), AIChE Journal, 18(3), 553-560. Tan S.H., 1976, Flare system design simplified, Hydrocarb Process, 46, 172-176. Vanderline L.G., 1974, Smokeless flares, Hydrocarbon Process, 53, 99-104.