A planar temperature imaging diagnostic has been developed and applied to an investigation of naturally occurring thermal stratification in an HCCI engine. Natural thermal stratification is critical for high-load HCCI operation because it slows the combustion heat release; however, little is known about its development or distribution. A tracer-based single-line PLIF imaging technique was selected for its good precision and simplicity. Temperature-map images were derived from the PLIF images, based on the temperature sensitivity of the fluorescence signal of the toluene tracer added to the fuel. A well premixed intake charge assured that variations in fuel/air mixture did not affect the signal. Measurements were made in a single-cylinder optically accessible HCCI research engine (displacement = 0.98 liters) at a typical 1200 rpm operating condition. Since natural thermal stratification develops prior to autoignition, all measurements were made for motored operation. Calibrations were performed in-situ, by varying the intake temperature and pressure over a wide range. Although the absolute accuracy is limited by the pressure-derived temperatures used for calibration, an uncertainty analysis shows that the precision of the diagnostic for determining temperature variations at a given condition is very good.
Application of the diagnostic provided temperature-map images that showed a progressive development of natural thermal stratification in the bulk gas through the latter compression stroke and early expansion strokes. Applying a PDF analysis with corrections for measurement uncertainties provided additional quantitative results. The data show a clear trend of going from virtually no stratification at 305° CA (55° bTDC), to significant inhomogeneities at TDC. Near TDC, the images show distinct hotter and colder pockets with a turbulent structure. Images were also acquired across the charge from the mid-plane to outer boundary layer at 330° CA and TDC. They show an increase in thermal stratification and a change of its structure in the outer boundary layer, and they provide a measure of the boundary-layer thickness. Where possible, results were compared with previous fired-engine and modeling data, and good agreement was found.