Numerical investigation on combustion and NO x emissions of a CH 4 / kerosene afterburner

. Cogeneration groups equipped with gas turbines usually operate with natural gas. The new requirements regarding the flexibility of functioning both with conventional and alternative fuels have led to the development of new solutions. An afterburning installation facilities, on the one hand, the combustion of fuels which can not normally be used in gas turbines, and on the other hand, allows a rapid adaptation of the cogeneration group to the user's variable thermal energy requirements. In some cases, an afterburning installation may contribute to the reduction of NOx emissions. It also increases the group's adaptability to changing environmental conditions, especially during the summer. Having in mind these tendencies and starting from the data collected from the 2xST 18 – Suplacu de Barcau cogeneration plant, this paper presents a numerical analysis of an afterburner in order to switch from natural gas to natural gas / kerosene functioning, with minimal modifications to the existing solution. A 3D steady RANS numerical integration of the Navier-Stokes equations has been carried out using the commercial software ANSYS CFX. For future research work, it will be taken into consideration the use of more complex reaction mechanisms and the variation of the spray angle will be considered. Also, in order to validate the CFD results a combustion experiments campaign will be conducted .


Introduction
Cogeneration groups equipped with modern gas turbines have to meet the requirements regarding the flexibility of functioning both with conventional and alternative fuels. Also, the pollutant emissions, have to be kept at a low level, even in partial load operation conditions.
The high oxygen content in the exhaust gases coming from the gas turbine, around 11-16 % vol. [1], can be used in the combustion of an additional amount of fuel (afterburning).
Afterburning is generally used on supersonic military aircrafts at take-off or during combat [2].
The utilization of an afterburning installation on cogeneration groups equipped with gas turbines increases the flexibility regarding of the used fuel and allows a rapid adaptation to the user's variable thermal agent requirements [3].
The control of the combustion process and of the afferent pollutants emissions, while complying with increasingly restrictive legislative requirements, remains one of the main challenges of designing a combustion system. In the field of cogeneration groups with gas turbines, combustion process modelling inside the afterburner becomes an integral part of the entire group design.
In the scientific literature, afterburning numerical modelling addresses, besides pollutant emissions reduction issues, aspects related to efficiency gains, compensation of environmental parameters variations, impact on the recovery boiler, optimization of flow conditions etc.
Sullivan-Lewis [4] analyzes several afterburner geometrical configurations regarding the functioning with low NO x and CO emissions, at a given fuel/air ratio and the geometry's influence on the combustion stability. Ganjehkaviri [5] has conducted a study on the impact of introducing an afterburner into a combined cycle power plant. Elmegaard et al. [6] studied the possibility of using a polycarbonate afterburning installation -functioning on natural gas / biogas or pyrolysis gas -to increase the efficiency of a power plant with indirect combustion gas turbine. Mehrabani et al. [7] proposed using an afterburning installation to compensate for the environmental parameters variation and to increase the efficiency of combined cycle power groups. Backlund [8] shows that parallel flow burners lead to: larger steam flows, a high efficiency of the gas turbine power plant, an efficient use of the oxygen from the exhaust gases coming from the gas turbine, a larger range of operating regimes, low NO x and CO emissions. Ahamed [9] analyzes various methods to improve the blending of fuel with exhaust gases from a turbo-engine, as well as the impact on the recovery boiler. Shi [10] optimizes the velocity and temperature fields upstream and downstream of the afterburner using numerical modelling. To determine the performance of an afterburner, Lezsovits et al. [11] used the commercial software ANSYS CFX 5.6.
The construction of the 2xST 18 -Suplacu de Barcau cogeneration plant, which has an afterburner installation functioning on natural gas with multiple parallel flows, led to researches regarding performance improvement. The afterburner was designed to function using as oxidant exhaust gases from a turbo-engine. However, in emergency cases (when there is a problem with the turbo-engine), it can also function with pure air as oxidant, air delivered by a fan.
Similar to the afterburner in the 2xST 18 cogeneration plant, a test bench for a 3modules afterburner with a single gas supply ramp has been built in INCDT COMOTI [12] ( Figure 1 a)).
Each afterburner module is composed of a deflector and 3 fuel nozzles. The central fuel nozzle, with a diameter of 6 mm, is the primary one. The other two are secondary ones with a role in flame propagation from one module to another. The deflector is fixed to the fuel ramp by the means of the 2 secondary fuel nozzles. The deflector (Figure 1 b)) is composed of a horizontal swirler plate and two vertical plates with the role of limiting the flame.
This configuration of the afterburner module insures a good mixing of the fuel (natural gas) with the oxidizer (exhaust gases from the turbo-engine / pure air). It insures the cooling of the area surrounding the afterburner, the stabilization of the flame front and leads to obtaining higher temperatures downstream, with low pollutants levels. The test bench has the possibility of testing the afterburner using as oxidant either exhaust gases coming from a turbo-engine or pure air ( Figure 2).
The test bench is mainly composed of: a TA2 turbo-engine running on natural gas (this is a TV2-117A turbo-engine converted by researchers from COMOTI to function with natural gas instead of kerosene), positioned horizontally, the afterburner and afterburning chamber positioned vertically in the turbo-engines exhaust chimney. The exhaust gases temperature and composition are monitored at a height of 1895 mm from the afterburner.  The proposed solution consists of modifying the central module of the afterburner presented in Figure 1 in order to function on kerosene, the adjacent natural gas modules remaining unchanged.
To better understand the implications of the proposed solution, there have been conducted first numerical simulations with natural gas as fuel and exhaust gases/ pure air as oxidant, for the configuration presented in F

CFD settings
A three-dimensional steady RANS numerical integration of the Navier-Stokes equations has been carried out using the commercial software ANSYS CFX.
In the numerical simulations it has been considered that the natural gas is composed only of methane gas (CH 4 ).
The methane gas mass flow corresponding to one module has been chosen in accordance with the design parameters of the afterburner from the 2xST 18 cogeneration plant, functioning on natural gas [12]. It has been considered that the total fuel mass flow is equally distributed to each module.
The kerosene mass flow, when the afterburner functions on fresh air (emergency regime), was calculated taking into account the maximum allowed thermal power for an afterburner module (350 kW) and the lower calorific value of kerosene [15].
In practice, the measurement of the flue gas temperature is performed in the centre of the afterburner chamber, on the monitoring plane positioned at 1895 mm above the burner, using the gas analyzer's integrated temperature probe. The flue gas reference temperature, resulting from the cogeneration plant design data, is 1043 K [12]. Preliminary experiments on natural gas, performed in COMOTI (Fig. 5) [16], showed a strongly turbulent flame. Thus in the numerical simulations the average temperature of the flue gases on the monitoring plane was compared with the reference temperature of 1043 K.
The fuel mass flow being thus set, the oxidant mass flow has been determined by varying it until an average temperature on the monitoring plan of 1043K, with an error of ± 1%, has been obtained.  Table 2 are presented the boundary conditions for the considered cases. The outlet has been considered of opening type, setting the pressure to be the atmospheric one.
In these numerical simulations an Eddy Dissipation combustion Model (EDM) and a k-ε turbulence model have been used. The combustion model was based on a two steps methane-air reaction mechanism, respectively a two steps kerosene-air reaction mechanism, both imported from the ANSYS library [17]. Also, both used reaction mechanisms take into consideration the formation of thermal NO and prompt NO.
In the case of functioning with kerosene, the fuel was uniformly injected into the atmosphere through the atomizer exit nozzle, under a total spray cone angle of 45°. The initial fuel droplets diameter was set at 0.1 mm. The break-up of the fuel droplets into smaller droplets was simulated using the CAB second break-up model from the ANSYS library. To simulate the transformation of the Jet A (kerosene) droplets into Jet A vapours a Liquid Evaporation Model from the ANSYS library was used.

Results and discussions
In Fig. 6 a) The trend followed by the NO mass fraction in the exhaust gases is consistent with the behaviour of the exhaust gases temperature. The NO concentration increasing with the rise of the temperature. The NO concentration is higher when fresh air is used as oxidant, both when CH 4 was used as fuel, as well as when kerosene was used as fuel. The high exhaust gases temperature favours the dissociation of the nitrogen molecules, which leads to the formation of nitrogen oxides [18].
In Fig. 8 and 9 are presented the temperature and NO mass fraction variations, along the central axis of the afterburner combustion chamber. As was expected the NO mass fraction increases with the temperature rise. The NOx emissions are significantly higher in the functioning regime CH 4 -air compared with the functioning regime CH 4 -exhaust gases. This confirms that the afterburner has been optimized to function in CH 4 -exhaust gases regime.
In Fig. 10 and 11 present the temperature variation on the transversal axis on the monitoring plan (positioned at 1895 mm from the afterburner).   Fig. 10 and 11 reconfirm the normal dependence between the NO emissions and the gases temperature. The lowest NO emissions are obtained for the functioning regime CH 4exhaust gases. This is consistent with the fact that the afterburner was designed to function optimally in this regime. The highest NO x emissions are obtained when operating on CH 4air, as emergency regime. This is observed also in the results presented in Fig. 9.
Regarding the functioning on kerosene, further numerical investigations are necessary in order to optimize the afterburner. Interestingly, NOx emissions values, when functioning on kerosene -air, are comparable to the values obtained for functioning on CH 4 -exhaust gases from the turbo-engine.
In Table 3 are presented the averaged temperature values obtained through numerical simulations, on the monitoring plan where the gas analyser probe is positioned (see Fig. 2). From the results in Table 3, it can be seen that, for the considered numerical simulation conditions, an error under 1 % for the average temperature on the monitoring plan has been obtained.

Conclusions
In order to study the possibility of switching the afterburner from functioning on gaseous fuel to liquid fuel (kerosene) functioning, a CFD analysis has been carried out. During all numerical simulations, it has been intended that the average temperature on the monitoring plan to be 1043 K ± 1%.
The numerical modelling, confirmed that the constructive afterburner solution has been optimized for functioning on gaseous fuel with exhaust gases from the turbo-engine, the lowest NO x emissions been obtained for this functioning regime.
Theoretical data and preliminary experiments have highlighted the turbulent structure of the flame, imprinted by the afterburner's modules' swirls, and the flame stability in the range of the analyzed CH 4 mass flow (0.006 -0.014 kg/s). In the case of functioning on kerosene, the flame stability and the low NOx emissions will have to be confirmed through experiments. The transition to kerosene functioning will be based on an original technical solution, which will be the subject of a patent application.
For future research work it will be taken into consideration the possibility of the afterburner functioning simultaneous with CH 4 and kerosene. Also, the use of more complex reaction mechanisms and the variation of the spray angle will be considered. In order to validate the CFD results, in parallel with the numerical simulations the combustion experiments campaign will be continued.