Quantitative time series for minor -species concentrations: Measurements and modeling in turbulent nonpremixed flames
Abstract
Combustion is utilized for over 80% of the energy consumed in the U.S. and occurs in the presence of turbulence in most cases. Fluctuations of scalars in turbulent flows often extend over three to four orders of magnitude in frequency and display characteristic power spectral densities (PSDs). However, analyses and models of turbulent nonpremixed flames by existing techniques usually neglect an explicit consideration of the time scales of turbulent fluctuations. Picosecond time-resolved laser-induced fluorescence (PITLIF) has been developed specifically for time-series measurements of minor-species concentrations in turbulent flames. The effects of fluorescence lifetime must be addressed since the laser power for PITLIF is insufficient to saturate the molecular transition. In this thesis, the final development of PITLIF is presented, permitting measurements of time series for the fluorescence lifetime and integrated fluorescence signal at rates sufficient for examination of turbulence statistics. The first known quantitative time-series measurements of OH are presented for a series of H2/Ar jet flames. The effects of lifetime fluctuations are found to be negligible so that the fluorescence time series alone can be used to characterize the concentration fluctuations. The measured PSDs for OH are found to be functions of Reynolds number and location within the turbulent jet. However, the different PSDs collapse onto one curve when normalized by the measured integral time scales. Measurements of OH and CH concentrations via PITLIF are also presented for a series of H2/CH4/N2 jet flames. The integral time scales are characterized and found to depart from the traditional scaling for nonreacting jets, implying that the scalar fluctuations are affected by the flame. A stochastic time-series simulation technique is presented which predicts these scalar time scales when using laminar-flamelet state relationships. The predictions from the model agree with measurements of radial profiles for OH integral time scales. This model appears to be sufficient for prediction of scalar fluctuation rates as long as the mixture-fraction statistics are adequately characterized.
Degree
Ph.D.
Advisors
King, Purdue University.
Subject Area
Mechanical engineering|Chemical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.