A New Air-Assisted Flare Tip Design for Managing Gas Flare Emissions (CFD Analysis)

Abstract

Recently, flares have been considered as a major source of air pollution from the petroleum refining industry. The United Nations has instigated an international effort related to the management of flare emissions to reduce the global warming impact related to flaring. Eliminating or removing the need for gas flares is difficult because these devices are generally used as safety devices to allow the combustion of flammable gases in a controlled fashion which supports safe operation. However, reducing flaring is generally possible using well-designed, efficiently operated flare equipment. In general, flare performance can be enhanced following the API-521 methodology and using assist-media including air and steam to achieve smokeless operation. This present work will discuss flare emissions in the petroleum refining industry and a method to manage flare emissions. Moreover, this work will discuss flare combustion efficiency (CE) and destruction and removal efficiency (DRE) in terms of efficient flare operation. This work uses actual operating flare data, published previously, which will be used in this work together with the CFD Code C3d. This code, developed at the USDOE Sandia National Laboratory, is based on a standard LES methodology to conduct transient flare analysis and is used to simulate flare operation to estimate flame shape and emissions produced. In this work, a new air-assisted flare tip design which uses the Coanda effect to improve flare operation was analyzed. This new flare design reduces the emission rate and demonstrates the design’s effectiveness. The analysis considers a flare 39″ high and 6″ diameter in the center of a 4m x 4m x 4m domain. Boundary conditions assume no cross wind and an ambient temperature of 300 K. The initial condition is a hydrostatic pressure profile across the computational domain. In the air-assist simulation, stoichiometric ratio will be a variable, and therefore, more than one case was considered.

1. Introduction

In the petroleum industry, flares are used to burn gases that cannot be processed or stored due to technical, economic, and safety reasons. Generally, flares are used to burn either large volumes of gases during start-up or shut-down processes or small amounts of gases during routine operation [1]. Flares are used not only in the petroleum industry but also in other industries, such as for sewage digestion, ammonia fertilization, and coal gasification [1,2]. Recently, the carbon emissions produced from the flares’ operation have been considered a major contributor to climate change [1,3,4]. The undesirable by-products from the flares’ operation affect the local population and the environment through noise, thermal radiation, smoke, and (NOx, SOx, soot, CO₂ and CO) emissions. To reduce the impact of these by-products, flare design and operation should be according to standards [5].

Generally, the top oil-producing countries are the main contributors to the increase in flaring volume due to the increase in oil production rates. Figure 1 shows the rate of flaring volume and oil production worldwide. Moreover, Figure 2 shows the gas flaring volume in the top 10 countries between 2012 and 2023. According to these data, Iraq is the third largest flaring country after Russia and Iran. This is related directly to oil production rate and improper flare design and operation.

Table 1. Flaring volume, intensity, and oil production between 2016–2023 in Iraq [6].

Years	Flaring Volume (Million m3/year)	Flare Intensity (m3 per Barrel of Oil Produced)	Oil Production (1000 Barrels per Day)
2016	17,551.51	10.82	4443.52
2017	17,843.04	10.98	4454.69
2018	17,767.14	10.55	4612.70
2019	17,914.22	10.35	4740.62
2020	17,373.79	11.64	4088.22
2021	17,891.97	12.00	4084.82
2022	17,902.88	10.96	4474.17
2023	17,689.50	11.16	4341.41

shows the oil production and flaring volume and intensity in Iraq between 2016 and 2023.

A flare’s performance is usually measured using two important parameters: combustion efficiency (CE) and destruction and removal efficiency (DRE) [7,8,9,10,11,12]. The first parameter, CE, measures the amount of CO₂ produced due to the complete combustion of the fuel, while the second one, DRE, measures the amount of consumed original fuel converted to CO and CO₂ [13,14,15,16]. According to studies, a good DRE indicates flame stability [13,14,15,16,17]. Moreover, factors like tip velocity and gas heating value affect flame stability. Usually, flares produce emissions under unstable operation conditions [18,19]. Combustion efficiency is also affected by flame lift-off when the tip velocity exceeds the burning velocity of the flare gas. USEPA recommends flare operation with a tip velocity less than the maximum allowable tip velocity measured using tip diameter, the density of the flared vent gas and the air, and combustion zone gas composition. Generally, flame stability is affected by factors such as flammability limits, flame speed, crosswind speed, and ignition temperature [19,20,21]. Field observations observed high combustion efficiencies at high flow conditions [14,22], while low combustion efficiencies below 98–99% were noticed by some field measurements at low flow conditions [23,24,25,26,27]. Studies showed that several factors cause low combustion efficiencies at low flow conditions, for example, the velocity of vent gas and flammability of the vent gas mixture [28].

The principle of injecting a jet into the crossflow is observed in many applications: for example, combustion equipment, mixing tanks, quenching systems, and drying systems. The distribution of flow in the domain of crossflow was reported to be affected by many factors including jet geometry, jet exhaust velocity, and the characteristics of the crossflow and injecting fluids [29]. Both computational and experimental approaches have been used to study the effects of different factors on the flow properties of jet gas injected into crossflow [30,31,32,33,34].

The Coanda effect can be explained as “when a jet of fluid is passed over a curved surface, it bends to follow the surface, entraining large amounts of air as it does so”. In other words, the Coanda effect can be defined as the tendency of a fluid to attach to a nearby surface and stay attached to it even if the surface curves away from the direction of the initial jet [35,36,37,38]. The Coanda principle can be found in various natural and man-made examples (such as medicine, industrial processes, maritime technology, and aerodynamics) due to the improvement in turbulent levels and entrainment [37,38]. On the other hand, the major problem of the Coanda effect is the detachment of the jet flow due to high static pressure at the nozzle exit. Therefore, to prevent or delay this problem, the static pressure at the nozzle exhaust must be reduced. For example, using a convergent–divergent nozzle can significantly lower the static pressure and solve the problem [39,40]. In the petroleum industry, the Coanda principle has been used to improve flare operation by designing Coanda flares. Since the Coanda effect entrains large amounts of air, this kind of flare compared to other types produces higher combustion efficiency, better smokeless operation, and less thermal radiation. Furthermore, the Coanda principle has been used in aircraft for thrust vectoring for short take-off and landing [40,41,42,43].

This paper aims to study the options for managing the air pollution problem due to flaring activities in Iraq and worldwide. The main aim of this work is to test a new air-assisted flare tip design under different operation conditions to evaluate the effectiveness of the proposed design in managing and reducing air pollution rates from flaring activities. Furthermore, the purpose of this paper is to increase the CE and DRE of flare operation using the new flare tip design. The following parts will describe the CFD simulation code C3d (version 2-19-24) used in this work as well as the results obtained from simulating different cases.

2. Methodology

The methodology of this work includes a real flare gas composition from the field with CFD code C3d (version 2-19-24) to simulate the utility and air-assisted flare cases. To show the relation between flare gas flow rate and soot formation, three gas flow rates (0.3 kg/min, 0.36 kg/min, and 0.42 kg/min) have been simulated and monitored using utility flares. The highest gas flow rate from these tests has been used in the air-assisted flare simulations due to the high soot formation and other combustion products. The effect of assist air on combustion was studied by considering three air flow rates (5.7 kg/min, 6.6 kg/min, and 8.2 kg/min) with the same gas flow rate (0.42 kg/min). These tests considered the stoichiometric air–fuel ratio of the combustion.

2.1. Flare Tip Design

The current study tests a new air-assisted flare tip design to enhance the performance by improving the CE and DRE and reducing the soot formation and other greenhouse gases. The proposed idea includes 32 small pipes (0.84″ ID, 1″ OD, and 45° inclination) around the stack near the tip to inject air at high velocity toward the flame to achieve smokeless operation. The design also includes a curved surface (Coanda effect) in front of the air injectors to produce homogenous flow from all directions and entrain more air toward the flame. The dimensions of the curved surface are 3″ in height and 2″ in width, with arc and chord lengths of 3.72″ and 3.55″, respectively. Finally, the flare tip was installed on a pipe 36″ tall. Figure 3 shows the proposed new air-assisted flare design.

2.2. Case Study

This study used the flare gas composition of a utility flare from an oilfield in the Iraq/Kurdistan region. This flare gas has the following properties: an average molecular weight of 19.25 kg/kg mole, a density of 0.782 kg/m³, and a heating value (LHV) of 46,266 kJ/kg. The complete flare data and analysis can be found in Maaroof et al. [5]. The original flare hydraulic limits were changed with a flare of diameter and height 6″ and 39″, respectively. Moreover, the original gas flow rate was changed with different flow rates and cases and applied in the absence of crosswind.

Table 2. Case study gas composition.

Basis = 100 kg mole
Components	Mole (%)	Mole (kg mole)	MWt (kg/kg mole)	Mass (kg)	Mass (%)
C1	83.34	83.34	16	1333.44	0.693
C2	9.501	9.501	30	285.03	0.148
C3	3.391	3.391	44	149.204	0.078
C6+	0.232	0.232	86	19.952	0.010
CO2	1.713	1.713	44	75.372	0.039
H2S	1.816	1.816	34	61.744	0.032
N2	0.007	0.007	28	0.196	0.000
Total	100	100.00		1925	1

shows the gas composition used in this study.

2.3. Computational Fluid Dynamics (CFD) Simulation

In this study, C3d LES based on Computational Fluid Dynamics (CFD) code version 2-19-24 was used to simulate the utility and air-assisted flare cases. More than six cases were simulated to study the flare performance and air-assisted effect on operation, in addition to the mesh independence study. These include three cases for utility flares at different gas flow rates of 0.3 kg/min, 0.36 kg/min, and 0.42 kg/min. Furthermore, for the air-assisted cases, the highest gas flow rate from the utility flare was applied with three different airflow rates within the limit of the air/fuel stoichiometric ratio: 5.7 kg/min, 6.6 kg/min, and 8.2 kg/min.

2.3.1. CFD Code C3d

C3d is a CFD and heat transfer code used for solving a variety of fluid mechanics and heat transfer cases that include flare and fire analysis. This code was developed at the USDOE Sandia National Laboratory and based on a standard LES methodology to conduct transient flare analysis. Moreover, in flare operation, C3d was used widely to estimate the flame shape and emissions produced. Various optional sub-models have been included in the code to facilitate simulating deposition, radiation heat transfer, aerosol transport, chemical reactions, combustion, and material decomposition [44]. The code was used previously in many works [45,46]. Originally, the C3d code started as a CFD tool called ISIS-3d that has been validated and used for simulating pool fire to study the thermal performance of nuclear transport packages [47,48,49]. Previously, many works used the code for studying the performance of different types of flares including utility flares, air-assisted flares, and large multipoint ground flares [44,50,51,52]. This is because the combustion model in the code has been developed, expanded, and approved for testing typical flare gas including propane, methane, ethylene, ethane, propylene, and xylene.

Also, the code has been used for estimating the flame size and shape as well as predicting the smoking tendency and heat flux from the flame to the ground. Additionally, C3d has been used in multipoint ground flares for studying the effect of airflow through the surrounding wind fens on flame height and shape during maximum firing conditions. The code has also been used with multipoint ground flares for estimating the space between flare tips and rows to guarantee adequate airflow to each burner during operation to avoid operation with smoke during maximum firing conditions [51]. Finally, the code has been used to investigate the effect of using a discrete and continuous ignition system on operation [50].

2.3.2. LES vs. RANS CFD

Based on studies, flare CE may not be computed accurately using Reynolds-Averaged Navier-Stock (RANS) approximations in the traditional CFD simulation. This is because the vertical coherent structures in the flames produce large-scale mixing that is not easy to reduce to a steady-state condition. Additionally, the unsteady information including flame shape and instantaneous mixing cannot be accurately captured by time averaging the equations in the RANS. Industrial flares operate in turbulent flow conditions that include large scales of time and length. The smallest of these scales is defined by viscosity and the biggest is on the order of flare tip diameter. In general, mixing rates limit combustion because of non-premixed combustion. The detailed kinetic mechanism of chemical reactions includes many central reaction steps with hundreds of spices and a wide range of reaction time scales. In the exothermic nature of the chemical reaction, both convective and radiative heat transfer take place. These processes are closely coupled: for example, the turbulent mixing of air and combustion gas affects the chemical reactions. Through turbulence, the chemical reactions change density and the mixing intensity. This happens due to gas temperature changes with the chemical reactions as heat is produced. Resolving all the length and time scales in practical turbulent combustion applications is hard to achieve and mostly not easy even with supercomputers. Instead, important flame features can be captured by resolving large time and length scales responsible for governing dynamics and applying sub-grid scale models for more homogenous smaller scales. In short, transient flare gas combustion observation can be performed more accurately using LES simulation compared to RANS methods [53,54,55,56,57].

2.3.3. CFD Models

In this simulation, the governing equations are discretized using a finite volume approach with orthogonal Cartesian coordinates to make the discretization very similar to a finite difference approach. All vector quantities in a finite volume formulation such as momentum are defined at the cell interface and scalar variables such as pressure and temperature are defined at cell centers. The flow equations in the C3d code are solved using a compressible version of the pressure-based solution algorithm [44,58]. Furthermore, the Large Eddy Simulation (LES) formulation is used to model turbulence. In this simulation, air is assumed to be incompressible with slight changes in temperatures. Moreover, the momentum equation used in this simulation is solved using a conservative form of the momentum flux vector ($\rho u$).

2.3.4. Computational Domain, Flare Model, and Boundary Conditions

In this work, the computational domain size was 4 m in height, 4 m in length, and 4 m in width. The height of the domain was on the z-axis and started from −0.1 to 0 to specify a ground (dry sand) for the flare model, then from zero to 4 m. The length of the domain was on the x-axis and started from −2 m to +2 m. Finally, the width was defined on the y-axis and started from −2 m to +2 m. Both the x-axis and y-axis started from −2 m to +2 m to define a center to locate the flare model. The flare model used in all simulation cases (both utility and air-assisted flare) was built out of carbon steel using the C3d code modeler. The dimensions of the flare model were 39″ height, 6″ ID, and 17.32″ OD. Figure 4 shows the domain and flare mesh and size, and Figure 5 shows the solid and mesh view of the flare model. Figure 5d shows the flow of the air through the Coanda surface using the cold flow approach. This approach includes injecting only air through the stack and air pipes to show the effectiveness of the Coanda surface.

The total number of hexahedral cells for the utility flare and air-assisted flare were 512,000 (80 × 80 × 80) and 729,000 (90 × 90 × 90), respectively. In the air-assisted flare case, the mesh was very fine in the area between 0.8 m and 1 m with 171,500 hexahedral cells (70 × 70 × 35), see Figure 5c, to capture the flow from the air pipes correctly. The grid has been refined using a higher number of cells to better capture the changes in the results and to confirm that the 512,000 cells (utility flare) and 729,000 cells (air-assisted flare) simulations were sufficiently accurate. Figure 6 and Figure 7 are mesh independence studies for the soot formation in the utility flare (0.42 kg/min gas) and the air-assisted flare (0.42 kg/min gas and 5.7 kg/min air). Generally, according to these figures, the increase in cell number (refining the mesh) provides almost the same results in both cases.

In this paper, the boundary conditions included no crosswind blowing from the x-axis or y-axis toward the flare, hydrostatics pressure defined across the domain, and the flare exit as a 3-D pressure. Generally, three 1-D sub-grids were used in the simulation for the ground, flare wall, and flare tip. In the first sub-grid, dry sand was used as the material for the ground. In the second sub-grid, carbon steel was defined as the material for the stack. Finally, in the last sub-grid, the mass flux and temperature of the assist air were controlled for each case. Finally, the gas flow rate was controlled through the 3-D boundary conditions by injecting air at 305 K from the bottom of the stack at the specified rates. Moreover, the 3-D boundary conditions included defining air across the domain at 300 K.

2.3.5. Physical Model

In this work, fluid flow was simulated using the LES turbulence model. In the simulation, the energy equation has been used to observe heat radiation effects. Moreover, individual species equations have been solved to monitor the fuel concentration and distribution, intermediate species, soot, and combustion products. The combustion model has been applied to provide the sink and source terms for the species equation as a function of species concentrations, local gas temperature, and turbulent diffusivity. In the simulation, the flame emissivity has been predicted using several models as a function of molecular gas composition, soot volume fraction, flame shape, flame size, and combustion effluent temperature profile. Generally, these variables rely on solutions of the mass, momentum, energy, and species equations. The radiation flux from the flame to the ground has been predicted using the radiation transport model which is also used to provide sink and source terms for the energy equation for predicting the distribution of the flame temperature.

2.3.6. Chemical and Soot Model

A combination of Arrhenius and Eddy breakup reaction time scales have been used to define the rate of combustion equations.

$$t_{total}=t_{reaction}+t_{turb}=\left({\displaystyle \frac{1}{A_k{T}^b{e}^{-\left(\frac{T_A}{T}\right)}}}\right)+{\displaystyle \frac{C_{eb}{dx}^2}{{\epsilon }_{diff}}}$$

(1)

where (T) = a local gas temperature, (A_k) = a pre-exponential coefficient, (b) = a global exponent, (C_eb) = the eddy breakup scaling factor, (T_A) = activation temperature, (dx) = the characteristic cell size, (t_turb) = the turbulence time scale, and (ε_diff) = the eddy diffusivity from LES module [44].

Combustion chemistry consists of, first, primary fuel breakdown reactions which yield intermediate combustion products (C₂H₂, H₂, CH₄, Soot, and CO). Secondary reactions involve burning intermediate products and forming soot. Additionally, reactions include reforming reactions with OH radicals. Moreover, the oxidizing species are simplified as water vapor. To allow for soot formation, equilibrium reactions between acetylene, methane, hydrogen, sulfur, and hydrogen sulfide are included [44]. The primary fuel breakdown reactions in this study are shown below:

CH₄ + O₂ → H₂ + CO + H₂O              C1 Breakdown

C₂H₆ + O₂ → 2.5H₂O + 0.5C₂H₂ + CO          C2 Breakdown

C₃H₈ + 1.5O₂ → C₂H₂ + 2H₂O + CO + H₂         C3 Breakdown

C₆H₁₄ + 3.5O₂ → 5H₂O + 2C₂H₂ + 2CO         C6+ Breakdown

H₂S + O₂ → SO₂ +H₂                 H₂S Breakdown

The secondary reactions are as follows:

H₂ + 0.5O₂ → H₂O          H₂ combustion

C₂H₂ + 0.8O₂ → 1.6CO + H₂ + 0.02C₂₀      soot nucleation soot formation

C₂H₂ + 0.01C₂₀ → H₂ + 0.11C₂       soot growth by acetylene addition

CO + 0.5O₂ → CO₂ + H₂O        CO combustion

C₂₀ + 10O₂ → 20CO           soot combustion

C₂H₂ + 3H₂ → 2CH₄       acetylene decomposition

CH₄ + CH₄ → C₂H₂ + 3H₂       acetylene formation

0.5S₂ + H₂ → H₂S       sulfur reduction to hydrogen sulfide

H₂S → 1.5S₂ + H₂        hydrogen sulfide decomposition

H₂S + 0.5SO₂ → 0.75S₂ + H₂O      elemental sulfur formation

The reforming reactions are as follows:

C₂₀ + 20H₂O → 20CO + 20H₂        soot steam reforming

0.75S₂ + H₂O → H₂S + 0.5SO₂        sulfur steam reforming

A global Arrhenius rate mode is used for these reactions. The consumption of fuel, soot, and intermediate species are evaluated by

$${\displaystyle \frac{{df}_{R_i}}{dt}}=-C\left[\prod _i^N{f}^{{PA}_i}\right]T^b{e}^{-\left({\displaystyle \raisebox{1ex}{$T_A$}\!\left/ \!\raisebox{-1ex}{$T$}\right.}\right)}$$

(2)

where (N) is the number of reactants, (T_A) is the effective activation temperature, (f_Ri) is the moles of each reactant, (i) and (C) are the pre-exponential coefficient, and (b) is the temperature exponent.

2.4. Transient Calculation and Post Processing

To calibrate the air and gas mass flow rates through the small pipes and stack, the simulation should first be run for 100 timestep. Next, the simulation time and timestep are set to 10 s and 1,000,000 steps, respectively, to reach stability in the calculations. The net reaction energy source and timestep size for the utility and air-assisted flare case (see Figure 8) show the simulation reached stability around 2 s. The net reaction energy source consists of reaction power and time. Reaction power is the gas flow rate (kg/s) times heating value (MJ/kg), for example, a gas flow rate of 0.42 kg/min (0.007 kg/s) multiplied by the heating value of the gas 46.266 MJ/kg will give 0.322 MW reaction power. At the beginning of the simulation (start of combustion), a lot of gas combustion results in a peak in the line, but after a few seconds, the reaction power value drops and stabilizes indicating the simulation is stable. The simulation stability is checked by comparing the value of reaction power in the curve with the mathematical value calculated by multiplying the gas flow rate by the heating value. For the air-assisted flare case, the reaction power is slightly higher due to the increase in gas velocity because of thrust. The assist air will increase the gas velocity by decreasing the gas flow cross-sectional area. Moreover, the simulation stability can be seen in timestep size figures in which the simulation timestep size is almost stable after a few seconds of combustion. This is used with a net reaction energy source to check the simulation stability.

3. Results

Paraview version (5.11.2) was used to visualize the flame size and shape and to extract mass fractions of the combustion products and soot formation in the domain using the probe feature.

3.1. Flame Size and Shape

The CFD code C3d was used to provide estimations and imaging of the flame size and shape for both utility and air-assisted flare cases. Three gas flow rates were used in the utility flare, 0.3 kg/min, 0.36 kg/min, and 0.42 kg/min, to show the relation between case rate and emission quantity. Figure 9 shows the utility flare operation at three different gas rates.

These rates were chosen after testing many cases with different rates lower than these rates. Less than the minimum gas rate of 0.3 kg/min causes flame backflow issues. This issue can be seen clearly in Figure 10 which is the cross-sectional view of the utility flare operation using gas rates 0.09 kg/min gas, 0.162 kg/min gas, and 0.25 kg/min gas. In other words, the 0.3 kg/min gas can be considered as the minimum gas rate for the case study gas mixture. To explain this more, the gas mixture has a density of about 0.78 kg/m³, and the air density is about 1.225 kg/m³, this means, at a low flow rate, the flare operation will face a phenomenon called Rayleigh–Taylor instability. This issue happens when a lighter fluid tries to push heavier fluid [59]. Alhameedi et al. [60] also used a six-inch pipe but with propane as a flare gas and lower flow rates (0.045 kg/min, 0.09 kg/min and 0.22 kg/min) than what was used in this study. This is again because the propane density is about 1.808 kg/m³, which is higher than the air density; therefore, the low flow rate will not cause density instability and backflow issues.

The highest flow rate in the utility case was chosen to study the effect of the air-assisted flare on flame stability and combustion emissions since burning more gas will produce more emissions. Figure 11 compares the utility flare operation with the new air-assisted flare design using the same amount of gas, 0.42 kg/min, and three air rates: 5.7 kg/min, 6.6 kg/min, and 8.2 kg/min. Based on the flame size in this figure, as the air rate increases, the flame size decreases. The amount of air used in the air-assisted flare was calculated based on the stoichiometric air–fuel ratio. Since natural gas with more than one composition was used as a gas in this study, three rates of air were applied after calculating the mixture’s stoichiometric air–fuel ratio.

The assist air was added to the combustion to enhance the CE and DRE and decrease pollution. In this study, to avoid a reduction in the flare performance and an increase in the air pollution, only the required amount of air for combustion was considered to avoid over aeration operation. Most of the flare design references agreed on using assisted medium (steam or air) to promote turbulence for achieving better mixing and avoid smoking. However, studies showed that increasing the amount of assist medium above a certain limit will have a negative impact on the flare operation and will reduce the CE significantly, which in turns lead to producing smoke and emitting considerable amounts of organic compounds [61,62,63].

3.2. Combustion Products

The probe function in Paraview was used to extract data on combustion products in the plume to evaluate the utility flare operation and study the effectiveness of the new air-assisted flare. Two locations in the domain 3 m and 3.5 m near the plume were defined to locate the probes and record the data. The radius of each probe was 3″ to cover as much as possible of the plume. Moreover, each probe was recording the data in ten different locations around the probe to cover as much as possible of the plume from all directions. Figure 12 shows the probe locations and height in the plume.

Table 3. Utility flare cases average temp, CO, CO₂, and soot at both sampling locations.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	Temp (K)	CO (Mass%)	CO2 (Mass%)	Soot (Mass%)
Utility Flare (0.3 kg/min gas)	0	0	3	3″	1.71 × 103	7.00 × 10−2	2.80 × 10−1	2.10 × 10−4
	0	0	3.5	3″	1.68 × 103	5.00 × 10−2	2.60 × 10−1	1.70 × 10−4
Utility Flare (0.36 kg/min gas)	0	0	3	3″	1.64 × 103	5.00 × 10−2	2.70 × 10−1	2.20 × 10−4
	0	0	3.5	3″	1.63 × 103	6.00 × 10−2	2.50 × 10−1	3.10 × 10−4
Utility Flare (0.42 kg/min gas)	0	0	3	3″	1.72 × 103	5.00 × 10−2	2.20 × 10−1	3.10 × 10−4
	0	0	3.5	3″	1.61 × 103	3.00 × 10−2	2.40 × 10−1	4.20 × 10−4

and

Table 4. Utility flare cases average C1, C2, C3, and C6+ at both sampling locations.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	C1 (Mass%)	C2 (Mass%)	C3 (Mass%)	C6+ (Mass%)
Utility Flare (0.3 kg/min gas)	0	0	3	3″	3.10 × 10−2	9.10 × 10−3	4.70 × 10−3	6.00 × 10−4
	0	0	3.5	3″	2.10 × 10−2	5.30 × 10−3	3.80 × 10−3	3.00 × 10−4
Utility Flare (0.36 kg/min gas)	0	0	3	3″	2.80 × 10−2	7.40 × 10−3	3.90 × 10−3	5.00 × 10−4
	0	0	3.5	3″	3.20 × 10−2	7.80 × 10−3	4.10 × 10−3	1.00 × 10−4
Utility Flare (0.42 kg/min gas)	0	0	3	3″	2.50 × 10−2	7.10 × 10−3	3.60 × 10−3	5.00 × 10−4
	0	0	3.5	3″	3.20 × 10−2	2.50 × 10−3	1.40 × 10−3	1.00 × 10−4

show the locations, size of the probe, and the combustion products of the utility flare cases.

Table 5. Utility and air-assisted flare average temp, CO, CO₂, and soot at both sampling locations.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	Temp (K)	CO (Mass%)	CO2 (Mass%)	Soot (Mass%)
Utility Flare (0.42 kg/min gas)	0	0	3	3″	1.72 × 103	5.00 × 10−2	2.20 × 10−1	3.10 × 10−4
	0	0	3.5	3″	1.61 × 103	3.00 × 10−2	2.40 × 10−1	4.20 × 10−4
Air-assisted Flare (0.42 kg/min gas and 5.7 kg/min air)	0	0	3	3″	1.53 × 103	5.00 × 10−3	2.00 × 10−1	2.00 × 10−4
	0	0	3.5	3″	1.50 × 103	4.00 × 10−3	2.30 × 10−1	2.00 × 10−5
Air-assisted Flare (0.42 kg/min gas and 6.6 kg/min air)	0	0	3	3″	1.44 × 103	8.00 × 10−3	1.90 × 10−1	1.50 × 10−4
	0	0	3.5	3″	1.50 × 103	1.00 × 10−3	2.40 × 10−1	1.90 × 10−5
Air-assisted Flare (0.42 kg/min gas and 8.2 kg/min air)	0	0	3	3″	1.61 × 103	1.10 × 10−2	2.30 × 10−1	6.00 × 10−5
	0	0	3.5	3″	1.07 × 103	3.00 × 10−3	1.50 × 10−1	2.70 × 10−5

and

Table 6. Utility and air-assisted flare average C1, C2, C3, and C6+ at both sampling locations.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	C1 (Mass%)	C2 (Mass%)	C3 (Mass%)	C6+ (Mass%)
Utility Flare (0.42 kg/min gas)	0	0	3	3″	2.50 × 10−2	7.10 × 10−3	3.60 × 10−3	5.00 × 10−4
	0	0	3.5	3″	3.20 × 10−2	2.50 × 10−3	1.40 × 10−3	1.00 × 10−4
Air-assisted Flare (0.42 kg/min gas and 5.7 kg/min air)	0	0	3	3″	1.10 × 10−3	4.00 × 10−4	2.00 × 10−4	4.00 × 10−5
	0	0	3.5	3″	1.00 × 10−3	3.00 × 10−4	2.00 × 10−4	2.00 × 10−5
Air-assisted Flare (0.42 kg/min gas and 6.6 kg/min air)	0	0	3	3″	1.10 × 10−3	4.00 × 10−4	2.00 × 10−4	3.00 × 10−5
	0	0	3.5	3″	4.00 × 10−4	1.00 × 10−4	1.00 × 10−4	1.00 × 10−5
Air-assisted Flare (0.42 kg/min gas and 8.2 kg/min air)	0	0	3	3″	2.20 × 10−3	7.20 × 10−4	3.70 × 10−4	4.50 × 10−5
	0	0	3.5	3″	7.20 × 10−4	2.30 × 10−4	1.20 × 10−4	1.50 × 10−5

show the combustion products of the utility and air-assisted flare cases.

Generally, these tables indicate that the combustion products are increasing with increasing flare gas rate. The air-assisted flare results show a decrease in the combustion product rates, indicating an improvement in flare operation and reduction in air pollution rates. This can be explained due to the effect of assist air on combustion in which the used amount of air was enough to combust more gas and reduce the air pollution.

3.3. Soot Formation

The soot formation in both utility and air-assisted flares was studied and demonstrated using Paraview and the CFD code. The soot formation in the utility cases shows that as the gas rate increases, the soot formation increases, too. This is because more fuel will need more air to achieve complete combustion and with the lack of the required air, part of the fuel will convert to soot. Figure 13 shows the soot (C20) formation in the three utility flare cases.

The effectiveness of the new air-assisted flare tip design in reducing environmental pollution appears in the change of the soot formation rate compared to the utility flare. Figure 14 shows the soot formation in the utility and the new air-assisted flare. As mentioned before, to show clearly the effectiveness of the new tip in controlling the pollution, the highest gas rate in the utility flare was used in the air-assisted flare case simulations. The differences between the utility and air-assisted flare soot formation are clear because the amount of air introduced in the assisted flare is enough for burning almost all fuel and reducing the smoke rate. The new flare tip design showed the ability to direct the air flow toward the flame efficiently. Moreover, the Coanda effect represented by the curved surface in the tip can keep the air attached to the surface and ensure there is air from all directions. Finally, the Coanda effect will make the air flow flatten, and therefore, it will be distributed equally from all sides around the flame.

Figure 15 and Figure 16 compare the soot formation of the utility and air-assisted flares at 3 m and 3.5 m sampling locations. The soot formation results show clearly the effect of the new air-assisted flare tip on managing pollution and improving the flare operation. At both sampling locations, the soot formation decreased significantly after using the new air-assisted tip. This improves the efficiency of the new design in managing and controlling the emissions produced by the flares’ operation.

3.4. Combustion Efficiency (CE) and Destruction and Removal Efficiency (DRE)

The performance of the utility and air-assisted flares was studied and compared using CE and DRE since these two factors determine the flare combustion quality. The DRE was calculated for the gas mixture and methane since the main compound in the gas mixture is methane. The equations below have been used to calculate these factors.

$$\mathrm{\%}\mathrm{C}\mathrm{E}={\displaystyle \frac{{\mathrm{C}\mathrm{O}}_2\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{p}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\right)}{{\mathrm{C}\mathrm{O}}_2+\mathrm{C}\mathrm{O}+\mathrm{C}1+\mathrm{C}2+\mathrm{C}3+{\mathrm{C}6}^{+}+\mathrm{S}\mathrm{o}\mathrm{o}\mathrm{t}\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{p}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\right)}}$$

(3)

$$\mathrm{\%}\mathrm{D}\mathrm{R}\mathrm{E}=1-{\displaystyle \frac{\mathrm{C}1+\mathrm{C}2+\mathrm{C}3+{\mathrm{C}6}^{+}\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{p}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\right)}{\mathrm{C}1+\mathrm{C}2+\mathrm{C}3+{\mathrm{C}6}^{+}\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}\right)}}$$

(4)

$$\mathrm{\%}{\mathrm{D}\mathrm{R}\mathrm{E}}_{\mathrm{C}\mathrm{H}4}=1-{\displaystyle \frac{\mathrm{C}1\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{p}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\right)}{\mathrm{C}1\left(\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{s}\mathrm{\%}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}\right)}}$$

(5)

Table 7. CE and DRE of the utility flare cases.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	CE (%)	DRE (%)	DRECH4 (%)
Utility Flare (0.3 kg/min gas)	0	0	3	3″	70.8%	95.1%	95.5%
	0	0	3.5	3″	76.6%	96.8%	97.0%
Utility Flare (0.36 kg/min gas)	0	0	3	3″	75.0%	95.7%	96.0%
	0	0	3.5	3″	70.6%	95.3%	95.4%
Utility Flare (0.42 kg/min gas)	0	0	3	3″	71.8%	96.1%	96.4%
	0	0	3.5	3″	78.3%	96.1%	95.4%

and

Table 8. CE and DRE of the utility and air-assisted flares.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	CE (%)	DRE (%)	DRECH4 (%)
Utility Flare (0.42 kg/min gas)	0	0	3	3″	71.8%	96.1%	96.4%
	0	0	3.5	3″	78.3%	96.1%	95.4%
Air-assisted Flare (0.42 kg/min gas and 5.7 kg/min air)	0	0	3	3″	96.6%	99.8%	99.84%
	0	0	3.5	3″	97.6%	99.9%	99.86%
Air-assisted Flare (0.42 kg/min gas and 6.6 kg/min air)	0	0	3	3″	95.1%	99.8%	99.84%
	0	0	3.5	3″	99.3%	99.9%	99.94%
Air-assisted Flare (0.42 kg/min gas and 8.2 kg/min air)	0	0	3	3″	94.1%	99.7%	99.68%
	0	0	3.5	3″	97.3%	99.9%	99.90%

show the CE, DRE, and DRE_CH4 for the utility and air-assisted flare cases at the two sampling locations. Generally, the DRE and DRE_CH4 results of the air-assisted flare are higher than the utility flare as more air will consume (destroy) more gas. The effect of the new air-assisted flare tip on the DRE and DRE_CH4 is shown clearly in Figure 17, Figure 18, Figure 19 and Figure 20. Regarding the CE, the air-assisted flare considerably improved the CE from about 70% in the utility flare to over 95%. The difference between the utility and the new air-assisted flares’ CE appears clearly in Figure 21 and Figure 22.

3.5. Soot Formation, CE, and Fuel Carbon Balance

To study more the effect of the new air-assisted flare on managing emissions and improving flare operation, the change in soot formation and CE were investigated and compared for both utility and air-assisted flares.

Table 9. Soot formation and CE of the utility and air-assisted flares.

	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	Soot (Mass%)	CE %
Utility Flare (0.42 kg/min gas)	0	0	3	3″	3.10 × 10−4	71.8%
	0	0	3.5	3″	4.20 × 10−4	78.3%
Air-assisted (0.42 kg/min gas and 5.7 kg/min air)	0	0	3	3″	2.00 × 10−4	96.6%
	0	0	3.5	3″	2.00 × 10−5	97.6%
Air-assisted (0.42 kg/min gas and 6.6 kg/min air)	0	0	3	3″	1.50 × 10−4	95.1%
	0	0	3.5	3″	1.90 × 10−5	99.3%
Air-assisted (0.42 kg/min gas and 8.2 kg/min air)	0	0	3	3″	6.00 × 10−5	94.1%
	0	0	3.5	3″	2.70 × 10−5	97.3%

shows the soot formation and CE for both flares (utility and air-assisted) at the defined sampling locations.

The fuel carbon balance shows the quantity of carbon removed from the fuel after combustion. This study compares the amount of C1, C2, C3, and C6+ in the fuel and the plume to understand the combustion more clearly. The original mass fraction of C1, C2, C3, and C6+ in the fuel compared to the plume shows a decrease in the amount of these compounds. Generally, the air-assisted flare removed more C1, C2, C3, and C6+ than the utility flare. This explains the DRE and combustion behavior of the air-assisted flare.

Table 10. Carbon balance study for the utility and air-assisted flares.

					C1 (Mass%)	C2 (Mass%)	C3 (Mass%)	C6+ (Mass%)	Total (Mass%)	Removed Carbon (Mass%)
					Fuel
					6.93 × 10−1	1.48 × 10−1	7.80 × 10−2	1.00 × 10−2	9.29 × 10−1
	x-Axis (m)	y-Axis (m)	z-Axis (m)	Radius (inch)	Plume
Utility Flare (0.42 kg/min gas)	0	0	3	3″	2.50 × 10−2	7.10 × 10−3	3.60 × 10−3	5.00 × 10−4	3.62 × 10−2	8.92 × 10−1
	0	0	3.5	3″	3.20 × 10−2	2.50 × 10−3	1.40 × 10−3	1.00 × 10−4	3.60 × 10−2	8.93 × 10−1
Air-assisted Flare (0.42 kg/min gas and 5.7 kg/min air)	0	0	3	3″	1.10 × 10−3	4.00 × 10−4	2.00 × 10−4	4.00 × 10−5	1.74 × 10−3	9.27 × 10−1
	0	0	3.5	3″	1.00 × 10−3	3.00 × 10−4	2.00 × 10−4	2.00 × 10−5	1.52 × 10−3	9.27 × 10−1
Air-assisted Flare (0.42 kg/min gas and 6.6 kg/min air)	0	0	3	3″	1.10 × 10−3	4.00 × 10−4	2.00 × 10−4	3.00 × 10−5	1.73 × 10−3	9.27 × 10−1
	0	0	3.5	3″	4.00 × 10−4	1.00 × 10−4	1.00 × 10−4	1.00 × 10−5	6.10 × 10−4	9.28 × 10−1
Air-assisted Flare (0.42 kg/min gas and 8.2 kg/min air)	0	0	3	3″	2.20 × 10−3	7.20 × 10−4	3.70 × 10−4	4.50 × 10−5	3.34 × 10−3	9.26 × 10−1
	0	0	3.5	3″	7.20 × 10−4	2.30 × 10−4	1.20 × 10−4	1.50 × 10−5	1.09 × 10−3	9.28 × 10−1

shows the mass fractions of C1, C2, C3, and C6+ in the fuel and the plume and the amount of removed carbon due to combustion.

4. Conclusions

This paper studied gas flare emission management using a new air-assisted flare tip design. The CFD code C3d was used to simulate utility and air-assisted flare cases at different flow rates. In this study, a utility flare gas composition from one of the oilfields in Iraq’s Kurdistan region was used as a flare gas in both utility and air-assisted flare cases. Moreover, flare hydraulic dimensions of 39” height and 6” inside diameter were used in this work. The new tip design included 32 small pipes of 1” ID and 45° inclination distributed equally around the tip to inject air from all directions toward the flame. Moreover, the tip consisted of a curved surface to provide a Coanda effect to the air injected from the pipes. To understand the effect of the new tip design on managing and reducing gas flare emissions, first, several utility flare cases were simulated using three gas flow rates to show the relation between the amount of gas burned and emission rates. Later, the highest gas flow rate was used in the air-assisted flare simulation. After calculating the stoichiometric air–fuel ratio, three rates of air within the limits of stoichiometric were applied in the simulation. The soot formation pictures showed the new air-assisted flare tip design was able to considerably reduce the emissions. Also, the thermal images showed that the air-assisted flare was able to consume more fuel compared to the utility flare. The probe function in Paraview was used to capture the products of combustion in the plume at two locations in the domain. The soot formation of the utility flare was significantly higher than the air-assisted flare. Moreover, the performance of the new assisted flare tip design was studied by considering the CE and DRE for both utility and air-assisted flare. The results showed that the air-assisted flare improved the CE from around 70% to over 95%. Generally, the results showed that the new tip design was able to reduce and manage the gas emissions produced by the gas flare. Based on results obtained, this flare tip design can be used for managing air pollution from flaring in oil and gas industries not only in Iraq but also worldwide. Moreover, this design can be used also as a steam-assisted flare in countries that have excess amounts of water.