A new optical diagnostic technique for burned gas temperature measurements was introduced using visible thermally-excited fluorescence of strontium monohydroxide (SrOH). The technique is a significant improvement over previously developed alkali metal-based techniques in that it requires only one tracer substance, strontium acetylacetonate in ethanol, compared to two or three alkali metal precursors. Combustion of the precursor forms 1.5 to 12 ppm SrOH, depending on equivalence ratio, and thermal excitation leads to visible light emission. For the purpose of temperature measurements, multiple emission bands were spectrally resolved and recorded with an intensified camera coupled to a spectrometer during experiments in an optical spark-ignited direct-injected engine. A hybrid experimental and computational approach was taken to validate the feasibility of SrOH based temperature measurements as well as to determine missing spectroscopic information. To that end, emission spectra were recorded for equivalence ratios ranging from lean to rich, and GT Power simulations provided a calibration base for burned gas temperatures. The optical engine was operated with early injection leading to near-homogenous mixing prior to ignition. It was confirmed that the measured data follow relationships consistent with a Boltzmann distribution of excited states populations. Furthermore, the experimentally determined energy difference between the excited [Formula: see text] and [Formula: see text] states of SrOH that form the basis of the temperature evaluation was found to be in good agreement with literature data. The calibrated analysis model then allowed to process spectra from over 400 single engine cycles from 10 runs with four different operating conditions to determine instantaneous burned gas temperatures. Average temperatures ranged from 1759 to 2490 K with standard deviations of 46–122 K. Variations were higher for more marginal fuel conditions. Self-absorption of the emission signals was characterized and would lead to a temperature error of no more than 0.5%.
Read full abstract