59029 - A Design of Experiments Based Investigation of the Influence of Hot Cross-Flow Gas on a Flox®-Based Single-Nozzle Liquid Burner 

Increased global demand for cleaner energy production and growing concern about using fossil fuels have urged many researchers to focus their work on developing more efficient and flexible combustion processes. In this regard, a FLOX^®-based liquid fuel single-nozzle burner is investigated for use in a Capstone C30 micro gas turbine (MGT). The main advantages of FLOX^®-based combustor systems are their decreased NO_x emissions and increased fuel flexibility. The C30 micro gas turbine operates at peak thermal power of 115 kW and generates an electrical load of about 30 kW. It consists of three burners arranged tangentially to an annular combustion chamber. In order to investigate the behavior of the FLOX^®-based liquid fuel burner regarding its combustion stability, exhaust gas quality, and flame characteristics under the influence of the other two burners, an atmospheric single-nozzle burner test rig is set up. The circulating gas in the annular combustion chamber is emulated by hot cross-flow gas produced by a 20-nozzle FLOX^®-based natural gas burner operated on a separate horizontal test rig. The variation and combination of the process parameters of both burners are done systematically according to Design of Experiments (DoE) as statistical design method. In order to perform an experiment most efficiently, a scientific approach to planning the experiment must be employed. DoE method is adopted rather than the conventional one-factor-at-a-time (OFAT) strategy, as DoE considers any possible interaction between the factors and reduces the number of experiments. Employing statistical design of experiments allows determining which input variables are responsible for the observed changes in the response, developing a model relating the response to the important input variables, and using this model for improving the combustor system. Corresponding information is extracted from data obtained during the experiments using response surface models such as Center Composite Design (CCD) and Box-Behnken Design. The models are then evaluated in terms of their accuracy of predicted response. The results are subsequently run through the Analysis of Variance (ANOVA) in order to allow for an objective conclusion about the effect of the factors on the selected responses which include thermal power (P_th) and global air equivalence ratio (ʎ_global) of both of the liquid and natural gas burners. The hot gas cross-flow interaction with the liquid fuel burner is assessed through analyzing exhaust gas emissions and averaged flame OH*-chemiluminescence images. The models developed by the DoE method are used to estimate the emissions and the flame geometrical properties of any other operating points that are not explicitly tested.