4.1.3. Reduction and control of NO_X emissions

Fundamental studies of individual droplet combustion in diffusion flames (Sarv et al. (1983)) indicate that low combustion temperatures associated with small (flame)/(droplet diameter) ratios are responsible for the low NO_X production detected in some instances. Spray combustion results in droplet interactions which decrease the local oxygen concentration around burning droplets, thus both flame temperature and NO_X are reduced.

Experiments have demonstrated the destruction of NO in hydrocarbon flames. Excess NO added to hydrocarbon flames is partially converted into HCN in the pre-flame (rich) zone, and the latter is transformed back into NO in amounts which depend on the flame stoichiometry.

A practical way to form less NO_X is the addition of water or steam (Gills (1973)) to the combustion environment, although at the expense of high running costs and loss of thermal efficiency (Bartok et al. (1971)) . The decrease in the formation of NO is caused by extra OH radicals produced by means of the reaction (Dryer (1977)):

H₂O + O OH + OH reac 34

which withdraws O radicals, thus slowing the Zeldovich mechanism.

Experiments in gas turbines (Singh et al. (1983)) with injection of water showed a large decrease of thermal-NO_X emissions by up to 85 % if the ratio (water/fuel) mass flow is between 0.01 and 0.4-0.5 (Shaw (1973)). In addition, other positive side-effects have been noted, like smoke reduction and power increase (Singh et al. (1983)). A similar effect is achieved by injection of steam, with NO_X reductions that can be ten-fold. However, steam does not affect the formation of smoke but increases emissions of CO and unburned hydrocarbons (Singh et al. (1983), Singh et al. (1972)).

The influence of water injection on fuel-N conversion to NO_X in gas turbines was also studied (Singh et al. (1983)). The conversion of fuel-bound nitrogen increases with higher (water/fuel) ratios, although the increase of conversion is more dramatic for water injected as a spray. It also has some influence on the conversion of NO to NO₂ as the ratio NO₂/NO is increased when the weight percentage of water in fuel is increased (Shaw (1973)).

Fuel modifications have also been attempted in gas turbines as a method for reducing NO_X. An exhaustive study upon the effect of additives (Shaw (1973)) demonstrates that soluble additives can act as heterogeneous catalysts efficient on NO_X reduction. Suitable additives are compounds of Mn, Fe, Co and Cu, whose effect is either to catalyse NO decomposition or to collect O ions. However, efficiency decreases as their concentration is increased. Other effective compounds are Na₂CO₃ (reducing NO_X by 20 %) and LiCO₃ (10 % reduction). Carbonates remove O ions by forming metal peroxides. NO_X depletion was also obtained with NaOH (17 % reduction) or hydrazine acetate (16 %).

Alternatively, fuel-water emulsions can be used to reduce NO_X (Singh et al. (1983)). In fact emulsions are more effective than water injection as experiments prove larger NO_X reductions with emulsions at similar (water/fuel) ratios. The reason seems to be the disruption caused by evaporation of the water, which decreases the size of the droplets formed.

One further method is to reduce the equivalence ratio of the combustion chamber. This can be achieved by recirculating the combustion products into the combustion air of boilers (Gills (1973)) , which causes reduction of the flame temperature. Thermal-NO is depleted, but not fuel-NO as its dependence on temperature is weak (Cunningham (1978)) . Experiments with vitiated air and fuel no. 2 oil or natural gas (Singh et al. (1972)) in gas turbines at an equivalent recirculation rate of about 26 % obtained a reduction of NO_X of 38 % and 30 % respectively. Recirculation of combustion gases also reduced the formation of CO and the emission of unburned hydrocarbons, although it has the drawback of increasing the amount of particles.

Conversely, leaning the primary zone of a combustor can be achieved by injecting increased amounts of air (White et al. (1982)) . Maximum reductions between 10 and 20 % can thus be obtained.

Minimisation of thermal and fuel-NO_X in gas turbines can be achieved by staging combustion (Hampartsoumian et al. (1991), Nimmo et al. (1991), White et al. (1982)) in separate zones with different equivalence ratios. Two combustor arrangements have been proposed:

1. Rich-lean configuration: A rich primary zone burns approx. 70 % of the fuel input, consuming most of the available oxygen. More secondary air is injected then, with reactions occurring at the limiting NO_X generation temperature. By this method fuel-NO_X is reduced. The emission of particles suffers an increase (Cunningham (1978)) as they are formed in the primary zone and they are more difficult to burn-out at a later stage.

2. Lean-lean configuration: Air and fuel are mixed and vaporised as much as possible before entering the lean primary zone. Low thermal-NO_X is achieved, but the yield of fuel-NO_X is higher.

CH + NO HCO + N	reac 35
CH + NO HCN + O	reac 36
CH + NO H + CO + N	reac 37

All are exothermic reactions. Shortcomings of this method are the likely emission of unburned fuel and efficiency loss.

Based on the reaction of nitric oxide with amine species, addition of ammonia to the combustion products (the SNCR process) at 1,000 °C can reduce NO (Muzio and Arand (1977)). NH₃ will form NH₂ radicals, which may undergo reaction with NO through the following channels (Halbgewachs et al. (1997)):

NH2 + NO HN2 + OH	reac 38
NH2 + NO N2 + H2O	reac 39

Finally, a reduction of the residence time of the combustion gases can be used as an alternative method. Thus N₂ and O₂ are allowed less time for reaction (Singh et al. (1983)). Reductions of up to 25 % have been achieved in gas turbines by moving air diluent holes upstream, thus shortening the flame length and the residence time.

Previous

Table of Contents

Next