Comparison of Two Methods to Predict Boundary Layer Flashback Limits of Turbulent Hydrogen-Air Jet Flames

Abstract

Flame flashback into the premixer is a serious issue in gas turbine combustion, especially for high hydrogen content fuels. Of particular interest is the risk of upstream flame propagation inside the wall boundary layer. Consequently, methods to predict the minimum flow velocities to prevent boundary layer flashback are sought by designers. In the first part of this paper two methods to predict boundary layer flashback limits are summarized and compared. The first method is a Damkohler correlation based on non-dimensional parameters developed at University of California Irvine (UCI). The correlation was developed based on the gathered experimental data at elevated pressures and temperatures (i.e. p = 3–7 bar, T u = 300–500 K, ϕ = 0.3–0.6) and successfully applied to a commercial gas turbine combustor. Due to its simplicity the Damkohler correlation is attractive for the design of gas turbine burners. But its applicability is limited to the turbulent combustion regime for which it was originally designed. The second method is called the “flame angle theory”. It was developed at Technische Universitat Munchen (TUM) and is based on a description of the physical process of boundary layer flashback. This method has been validated with experimental data at atmospheric pressure and a wide range of preheating temperatures and equivalence ratios (T u = 293–673 K, ϕ = 0.35–1.0). Since it describes the physical process of boundary layer flashback based on a set of sub-models it should be generally applicable to all operating conditions if the sub-models are appropriate. To verify this, the flame angle theory is applied to high pressure conditions in the second part of this paper. A comparison with results from the Damkohler correlation shows that the predicted flashback limits are in a reasonable range. However, the degree of agreement between Damkohler correlation and flame angle theory strongly depends on equivalence ratio because the Damkohler correlation does not account for the changing susceptibility of different hydrogen-air mixtures to flame stretch. For that reason, a modified Damkohler correlation has been derived at TUM from the flame angle theory and is presented in the third part of this paper. This correlation combines the advantages of the other two methods as it features high usability and is generally applicable to all operating conditions.