NOx-Abatement Potential of Lean-Premixed GT Combustors

Abstract

The influence of the structure of perfectly premixed flames on NOx formation is investigated theoretically. Since a network of reaction kinetics modules and model flames is used for this purpose, the results obtained are independent of specific burner geometries. Calculations are presented for a mixture temperature of 630 K, an adiabatic flame temperature of 1840 K, and 1 and 15 bars combustor pressure. In particular, the following effects are studied separately from each other: • molecular diffusion of temperature and species; • flame strain; • local quench in highly strained flames and subsequent reignition; • turbulent diffusion (no preferential diffusion); • small scale mixing (stirring) in the flame front. Either no relevant influence or an increase in NOx production over that of the one-dimensional laminar flame is found. As a consequence, besides the improvement of mixing quality, a future target for the development of low-NOx burners is to avoid excessive turbulent stirring in the flame front. Turbulent flames that exhibit locally and instantaneously near laminar structures (“flamelets”) appear to be optimal. Using the same methodology, the scope of the investigation is extended to lean-lean staging, since a higher NOx-abatement potential can be expected in principle. As long as the chemical reactions of the second stage take place in the boundary between the fresh mixture of the second stage and the combustion products from upstream, no advantage can be expected from lean-lean staging. Only if the primary burner exhibits much poorer mixing than the second stage can lean-lean staging be beneficial. In contrast, if full mixing between the two stages prior to afterburning can be achieved (lean-mix-lean technique), the combustor outlet temperature can in principle be increased somewhat without NO penalty. However, the complexity of such a system with a larger flame tube area to be cooled will increase the reaction zone temperatures, so that the full advantage cannot be realized in an engine. Of greater technical relevance is the potential of a lean-mixlean combustion system within an improved thermodynamic cycle. A reheat process with sequential combustion is perfectly suited for this purpose, since, first, the required low inlet temperature of the second stage is automatically generated after partial expansion in the high pressure turbine, second, the efficiency of the thermodynamic cycle has its maximum and, third, high exhaust temperatures are generated, which can drive a powerful Rankine cycle. The higher thermodynamic efficiency of this technique leads to an additional drop in NOx emissions per power produced.