Numerical measurements of two-staged gas turbine combustion for hydrogen enrichment of hydrocarbons

Problem: High operating pressures in gas turbine destroys the flame

To investigate the effects of initial temperature, addition of hydrogen and operating pressures above atmospheric in sequential increase up to 32 bar. The study considers three different configurations for mixing the quality of fresh mixtures: simple annular, radial jets, and lobe mixers. The best of three will be chosen to further investigate for low emissions.

A turbulent combustion model assumes that the flame in an explosion is a library of flamelets. The reaction model is handled in two steps: i) the reaction closure with the transport equations(s), reaction rate expression and other related expansive terms. The combustion model relies on the concept that the flame propagates into the unburned mixture with a defined velocity (which, of course, differs with respect to the chosen relative position, e.g. along the unburned or burned side of the flame brush); and ii) the reaction subclosure of the global quantity of premixed turbulent combustion, turbulent flame speed S_T. The S_T increases with decreasing Le and increasing preferential diffusion PD. Expectedly, lean H₂ mixtures are characterised by substantial differences in thermal diffusivity, and among mass diffusivities of governing fuel and oxidant, the effects (of Le and/or PD) are very strong even if the turbulence intensity is exceedingly higher than the unstretched laminar flame speed – a perspective of H₂ explosion modelling.

Substantiating the S_T model for ignition and time-evolution phenomena allow an understanding of and ability to locate the zone of ignition and the inception phase of flame propagation. Inclusion of the DL instabilities still furthers pictures out the flame instabilities, an important occurrence in H2 modelling following its high diffusive nature. Moreover, of course, the success of explosion simulation depends on the selection of the turbulence models, with LES retaining the edge over RANS in handling wake-generated turbulence. The basic physical concepts and terms, and the models that elucidate hydrogen combustion/explosion, serve as a common platform to address relevant issues in conventional combustion modelling and explosion modelling. The effects of molecular transport on S_T, mean the flame brush thickness and structure still need to be addressed in both cases, including the strong dependence of S_T on Le and PD (it is noted that it is hard to interpret their individual contributions). It is worth noting that these effects are most pronounced in H2–air mixtures than in HC mixtures.

In a developing flame (or explosion), pressure evolution is trapped, which requires that the reaction model is sensitive to pressure effects. The AFSW reaction model holds the advantage by invoking the high-pressure effects explicitly, which actuates the high pressure effects with the explicit terms, inverse of normalised molecular kinematic viscosity. However, this model needs to be indued for measurable quantities, such as maximum pressure, time to build the maximum pressure, duration of pressure pulse and so on from the experiments. In another approach, the reaction models based on the effects due to the Markstein number, flamelet quenching, or small-scale wrinkling provide no clue to predicting the leading-edge speed, a concept viable to H₂ explosion and consequent pressure build-up evaluation.

Earlier studies in collaboration with the then ALSTOM, Baden (Switzerland) have shown that the AFSW model was found to accurately predict the flame anchoring positions and yield very good quantitative results for various fuels, including diesel/air mixtures and operating pressures up to 32 bar (Aluri, N.K. and Muppala, S.P.R., 2020). The investigator has recently established a close research collaboration with a Japanese experimental Kido group and the Paul Scherrer Institute (Switzerland). Both groups study combustion characteristics in premixed turbulent flames.

A number of researchers have used the supervisor’s combustion model, known in the literature as Muppala’s combustion model.

Benefits of this study: this two-staged premixed combustion system gives low NOx emissions.

To summarise, this research will undertake numerical measurements of the characteristics of a two-staged gas turbine combustion system. The results will contribute to better thermal efficiency, fuel flexibility and lower emissions