In-Cylinder Temperature Estimation from an Optical Spray-Guided DISI Engine with Color-Ratio Pyrometry (CRP)

Abstract

<div class="htmlview paragraph">Color-ratio pyrometry (CRP) is a technique for estimating the temperature and loading of soot, based on its thermal emission spectrum. This technique is contrasted with conventional two-color pyrometry which requires absolute measurements of the radiation intensity, either at two specific wavelengths or ranges of wavelengths. CRP uses two ratios, obtained by measuring the radiation intensity for three wavelengths or wavelength bands. CRP has been implemented here by using a digital CCD camera, and full details of the calibration are reported. Because of uncertainties in the emissivity of reference sources (such as tungsten ribbon lamps, in which the emissivity depends on temperature and wavelength), then a spectroscopic calibration of the CCD camera has been used. Use of a CCD camera is not straightforward because of internal digital signal processing (DSP), so full details are given of the calibration and technique implementation. Calibration results, together with an error analysis, showed an accuracy of ±50K within the combustion temperature range and a preference for the temperature estimates based on the red/green ratio over the red/blue ratio at low temperatures and the red/blue ratio over the red/green ratio at high temperatures. Temperature estimates are shown to be insensitive to the concentration of soot.</div> <div class="htmlview paragraph">Combustion imaging was later carried out on a spray-guided direction injection spark ignition engine with optical access. Temperature estimates under various imaging timings and mixture strengths are presented. With a fixed operating point the maximum in-cylinder temperature occurred just after top center, and this coincided with the maximum rate of heat release. When the mixture strength was varied, the maximum temperature (at a particular crank angle) was rich of stoichiometric. Finally, temperature and <i>KL</i> results using the variable <i>KL</i> algorithm were reported together with discussions about soot formation.</div>