K. T. Rhee

Infrared spectral analysis of engine preflame emission

Abstract An investigation has been conducted to examine the previously observed infrared emission occurring at the end of an engines compression period. The infrared preflame spectrum is measured and quantitatively compared with a radiation emission model incorporating wall effects. The likelihood and degree of contributions to the observed emission by intermediate species under non-local thermodynamic equilibrium conditions are examined, and the viability of applying high-speed infrared imaging as a diagnostic technique for observing the distribution of fuel vapour is demonstrated. A Fastie—Ebert-type grating spectrophotometer was constructed to employ a 64×64 pixel platinum-silicide charge-coupled device imager. This instrument was integrated with a four-camera infrared spatial imaging system. The combined apparatus allows the simultaneous recording of the infrared spectra and four spatial images, each recorded through separate narrow-band filters. One filter is centred at 2.47 μm and another centred at 3.43 μm, covering strong H2O and hydrocarbon bands. Images are recorded at a framing rate of 1880 frames/s in 64-frame sequences from consecutive engine cycles and triggered at a chosen crank-angle position determined by a crankshaft-mounted encoder wheel. To interpret the measured spectrum, a line-by-line radiation model was created utilizing the high-resolution transmission (HITRAN) database of molecular parameters. This database includes spectral line data for radical and intermediate species including hydroxyl (OH) and formaldehyde (HCHO), and the alkyl compounds methane (CH4), ethylene (C2H4), and up to ethane (C2H6). The model predicts the spectra of an isothermal, homogeneous gas mixture consisting of fuel, residual gases H2O and CO2 in air at a given temperature, pressure and concentration, and considers the effect of interaction with an emitting combustion chamber wall. Preflame images and spectra were recorded from an optically accessible spark-ignition engine using four fuels representing minimal (CH4, C2H4), marginal (C2H6), and significant (propane, C3H8) cool-flame chemistry. Results show the measured preflame spectra of the four fuels to have similar large-scale features and magnitude, including the propane fuel known for chemiluminescence during low-temperature reactions. The preflame spectrum of the latter fuel mixture is notable for its lack of spectral features expected of HCHO or OH, suggesting that emission by intermediate species produced by low-temperature mechanisms is not readily evident in the infrared preflame spectrum. There is a high degree of correlation between the measured preflame spectra of CH4 and C2H4 mixtures and the spectrum modelled using parameter values reasonably representative of engine conditions. The prominent features of the observed infrared preflame spectra are explained by thermal emission of the species considered. Since reported absorption cross-sections of n-heptane, n-dodecane, and gasoline at 3.39 μm are greater than that of CH4, these fuels are expected to have thermal emissions of similar if not greater magnitude near this wavelength. Infrared images of fuel vapour during a diesel combustion cycle that have been recorded with the four-camera system are presented. Significantly, it has been shown that mid-infrared imaging can be used as a technique for observing in-cylinder fuel vapour distributions.

Computation of Radiation Heat Transfer in Diesel Combustion

A theoretical model of radiation heat transfer has been developed. A computation of radiation heat flux at a particular location in the combustion chamber by using the present model requires in-cylinder time- and space-resolved species data and cylinder pressure. From the species data, the burned fuel/air ratio distribution is inferred to compute space-resolved adiabatic flame temperature. For the computation of the spectral emissivity of an isothermal volume of adiabatic temperature containing soot, the Rayleigh-limit expression is used. The refraction indices in the expression are obtained by using the dispersion equations based on the electronic theory encompassing both free and bound electrons. For the spectral emissivity from the gaseous component in the volume, the semi-empirical band model is used. A parametric analysis of radiation heat transfer in diesel combustion is made by using the present model; a prediction by the model qualitatively compares with some of the reported experimental data.

Quantitative Imaging of In-Cylinder Processes by Multispectral Methods

Abstract : With the objective of achieving better investigation of engines-fuels by obtaining instantaneous quantitative imaging of in-cylinder processes, several steps have been taken for some years at Rutgers University. They are: (1) Construction of a new Multispectral high-speed infrared (IR) digital imaging system; (2) Development of spectrometric analysis methods; (3) Application of the above to real-world in-cylinder engine environments and simple flames. This paper reports some of results from these studies. The one-of-a-kind Rutgers IR imaging system was developed in order to simultaneously capture four geometrically (pixel-to-pixel) identical images in respective spectral bands of IR radiation issued from a combustion chamber at successive instants of time and high frame rates. In order to process the raw data gathered by this Rutgers system, three new spectrometric methods have been developed to date: (1) dual-band mapping method; (2) new band-ratio method; and (3) three-band iteration method. The former two methods were developed to obtain instantaneous distributions of temperature and water vapor concentrations, and the latter method is to simultaneously find those of temperature, water vapor and soot in gaseous mixtures, i.e., to achieve quantitative imaging. Applications of these techniques were made to both SI and CI engine combustion processes as well as bench-top burner flames. Discussion is made on the methods and new results.

Spectral IR Images of Direct-Injection Diesel Combustion by High-Pressure Fuel Injection,

Abstract : Instantaneous successive spectral infrared (IR) images were obtained from a spray plume in a direct injection (DI) type compression-ignition (Cl) engine during the compression and combustion periods. The engine eqwpped with a high pressure electronic-controlled fuel injector system was operated by using D-2 Diesel fuel. In the new imaging system used for the present study, four high-speed IR cameras (with respective band filters in front) were lined up to a single optical arrangement containing three spectral beam splitters to obtain four spectral images at once. Two band filters were used for imaging the water vapor distribution and another two band filters were placed for capturing images of combustion chamber wall or soot formation. The simultaneous imaging was successively triggered by signals from an encoder connected to the engine. The fuel injection parameters were precisely controlled and the pressure-time (p-t) history was obtained for individual sets of images. The start of fuel injection was varied through four different crank angle positions. Mentioning some results from the study, the spectral IR images had no resemblance with the ones obtained using a visible-range camera from a comparable engine system as reported by others. In general, the present spectral images taken at the same crank angle were not mutually comparable.

MTBE for Improved Diesel Combustion and Emissions

Reduced emissions from the spark-ignition engine, when fueled by gasoline containing small amounts of MTBE, have led us to explore similar positive results in compression-ignition (CI) engine combustion by adding this oxygenate compound to Diesel fuel. This study was performed in two separate laboratories by employing the respective experimental apparatus. When a pre-chamber type CI engine was operated by using Diesel fuel mixed with several volume portions of MTBE, including 5, 10 and 15%, several positive results were obtained, as compared with those from the baseline neat Diesel-fueled operations: (1) The engine delivers overall comparable or better performance characteristics: (2) The brake thermal efficiency is higher at the advanced and late injection times; (3) Some considerable reduction of both soot and NOx emissions is found; (4) The ignition delay increases but the combustion duration decreases. In order to help explain the unexpected responses of CI combustion to the above fuel modifications, high-speed in-cylinder spectral infrared digital imaging was performed for the same fuel modification. Some plausible and consistent results are observed, which help explain the above findings. 5 refs., 9 figs.

Investigation of a Direct Injection Diesel Engine by High-Speed Spectral IR Imaging and KIVA-II,

Abstract : In-cylinder process of a direct injection (DI) compression ignition (CI) engine was studied by using the Rutgers high-speed spectral infrared (IR) imaging system and the KIVA-II computer code. Comparison of the engine measurements with the computational prediction was attempted. In order to perform the instantaneous IR imaging, a Cummins 903 engine cylinder head was modified by installing an optical access in place of one of the intake valves, which required designing a new rocker-arm mechanism. The measurements obtained using the high-speed dual spectra IR imaging system were processed by the conventional two-color method which employed soot as the radiating target. The KIVA-II program was coded in order to match engine and operation conditions to those employed in the present measurements for achieving mutual consistency of the analysis. (MM)

INCOMPLETE COMBUSTION IN ONE-END-OPEN CREVICES

The present paper considers the processes of incomplete combustion in in-cylinder crevices with clearances slightly greater than quenching distance. For this, an experimental work has been carried out by using a premixed constant-volume combustion chamber. In the chamber, the propagation of flame through the combustible gas contained in individual crevices with various geometries was investigated by two means: high speed schlieren photography to obtain the idiosyncrasy of the in-crevice flame behavior; and fast-response thin-film thermocouples mounted flush with the crevice wall to measure the flame propagation speed, the instantaneous surface temperature, the instantaneous heat flux through the crevice wall, etc. From the investigation, the origins of unburned hydrocarbons formed in the in-cylinder crevices were surmised. In addition, an improved numerical method is presented for the computation of heat flux through a slab (here, the crevice wall) by using the timed surface temperature measured by the thin-film thermocouple.

Fuel Effects on Diesel Combustion Processes

Abstract : The crank angle locations for the first occurrences of several main combustion events in a Diesel engine were investigated for varied fuel parameters. The events studied include: preflame reactions; premixed flame propagation; start of pressure rise; maximum rate of pressure rise (dp/dt); and peak cylinder pressure. The fuels employed in the study were in two groups: (1) Base fuel-1 and derivatives prepared by mixing it with small doses of a cetane number (CN) enhancing additive; and (2) Base fuel-2 and those made by adding different amounts of bio-Diesel fuel. The experiment was performed by using a single-cylinder direct-injection (DI) Diesel engine equipped with an electronically controlled high-pressure fuel injection unit. The in-cylinder processes during the periods of ignition delay and combustion reaction were measured by using a high-speed multispectral infrared (IR) imaging system developed at Rutgers University. The other events were found from the pressure-time history. The purpose of using these fuels was to investigate: additive effects on the (invisible) preflame reaction and visible premixed flame development; flame behaviors of bio-Diesel fuels; CN effects on in-cylinder reactions; and others. There is some evidence that the formation of the visible flame kernels may not be directly related to the preflame reactions when the additive is used to increase CN. The reactions during the ignition delay of bio-Diesel fuels were rather unpredictable, therefore requiring additional investigation. Among the most indicative timelines for determining a fuels CN were those of: the maximum dp/dt; the start of pressure rise; the first premixed flame; and the peak pressure. The timeline of maximum dp/dt seems to be most insensitive to the variation of injection timing. Some new findings are also reported in the paper.

Flames and Liquid Fuel in an SI Engine Cylinder during Cold Start

Abstract : The flame propagations in the very first firing and subsequent cycles in an SI engine during cold start were studied to gain a better understanding of reaction fronts associated with liquid fuel (regular unleaded) in the cylinder. This work was performed using the Rutgers high-speed spectral infrared digital imaging system on a single-cylinder engine with optical access. The engine was mounted with a production engine head mated with a conventional pon fuel injection (PFI) system. In the study, four images in respective spectral bands were simultaneously obtained at successive instants of time, which was done for eight sequential cycles. This multiple-band successive-imaging was repeated in intervals of about two minutes over a period of more than twenty-five minutes after the engine start. During this experiment, the temperature changes at the intake port, the water jacket and the exhaust gas were monitored. In addition, pressure-time data was obtained from individual cycles in order to gain some insight into the overall in-cylinder reactions. The first firing cycle exhibited almost invariably weak flame propagation, which was followed by very intense flame fronts in the next cycle. Note that the flame propagation in the first cycle seems to only indicate consumption of the filel vapor available in the cycle. The flames in the third cycle were also intense in some cases, but mostly weaker than those in the second. Upon formation of the flame front in the beginning of combustion, some exceedingly strong local reactions started to grow, but no earlier than l5CA after TDC.