Download
1 / 26
290 likes | 586 Views
Part 2 Flame Observation and Measurement. In-Cylinder Pressure and Flame Measurement. Dr. Manoochehr Rashidi Engine Research Center Shiraz University. http://succ.shirazu.ac.ir/~motor/ motor@shirazu.ac.ir. Schematic arrangement of the transparent piston engine.
E N D
Part 2Flame Observationand Measurement In-Cylinder Pressure and Flame Measurement Dr. Manoochehr Rashidi Engine Research Center Shiraz University http://succ.shirazu.ac.ir/~motor/ motor@shirazu.ac.ir
Quartz piston assembly used for obtaining high speed motion picture of flame
HYCAM rotating prism camera, 10 000 frames/sec on 16 mm film • Object lens • Image size limiter • Segment shutter • 1st field lens • 1st prism • plane compensation prism 11 Image 12 Prism 13 Ocular 14 Synchronized gear drive 7 2nd field lens 8 2nd prism 9 Intermediate lens 10 U prism
Color photographs from high speed movie of spark ignition engine combustion process, taken through glass piston. Ignition timing 30 before TC, 1430 rpm.
Non uniform flame propagation is one of the causes of cyclic variation in engine
Microshadowgraphs of flame at various engine speeds showing effect of turbulence
Synchronization arrangement, for simultaneousflame front and pressure measurement.
Typical profile of flame front at various crank angleInner circle is the visible part of the combustion chamber
Entrained (or burning) velocityFull line is least square, and doted line is from model
Variation of flame geometry and velocity parameters during four individual combustion cycles. Variables shown in the figure are, flame radius rf, burned gas radius rb, normalized entrained volume yf, burned volume yb,. normalized flame front area af, laminar area aL, flame front speed ub, burning speed Sb, and laminar flame speed SL.
Figure in previous page shows results from an analysis of cylinder pressure data and the corresponding flame front location information (determined from high-speed movies through a window in the piston) of several individual engine operating cycles. The combustion chamber was a typical wedge design with a bore of 102 mm and a compression ratio of 7.9. The flame radius initially grows at a rate that increases with time and exhibits substantial cycle-by-cycle variation in its early development (Fig. a). Later (rf > 30 mm) the growth rate, which approximates the expansion speed ub , reaches an essentially constant value. The flame radius rf is initially equal to the burned gas radius rb it increases above rb as the flame grows and becomes increasingly distorted by the turbulent flow field (Fig. b). Eventually rf - rb goes to an essentially constant value of about 6 mm. This difference, is approximately half the thickness of the turbulent flame brush.
Normalized enflamed and burned volumes, and flame front area and laminar burning area, are shown in Fig. c. Volumes are normalized by the cylinder volume, and areas by Rh, where h is the average clearance height and Rthe cylinder radius. Discontinuities occur in the flame area af at the points where the flame front contacts first the piston face and then the near cylinder wall. The laminar area AL is initially close to the flame area Af and then increases rapidly as the flame grows beyond 10 mm in radius. During the rapid burning combustion phase (yf > 0.2) the value of yf is significantly greater than yb . During this phase, the laminar area exceeds the flame area by almost an order of magnitude. These observations indicate the existence of substantial pockets of unburned mixture behind the leading edge of the flame. The ratio of the volume of the unburned mixture within the turbulent flame zone, to the reaction-sheet area within the flame zone, defines a characteristic length lr , which can be thought of as the scale of the pockets of unburned mixture within the flame. lr is approximately constant and of order I mm.
These flame geometry results would be expected from the photographic observations of how the flame grows from a small approximately spherical smooth-surfaced kernel shortly after ignition to a highly wrinkled reaction-sheet turbulent flame of substantial overall thickness. Initially, the amount of unburned gas within the enflamed volume is small. During the rapid burning phase of the combustion process, however, a significant fraction (about 25 percent) of the gas entrained into the flame zone is unburned. The front expansion speed uf, burning speed Sb, and laminar flame speed SL are shown in Fig. d. The expansion speed increases as the flame develops to a maximum value that is several times the mean piston speed of 3.1 m/s and is comparable to the mean flow velocity through the inlet valve of 18 m/s. The burning speed increases steadily from a value close to the laminar flame speed at early times to almost an order of magnitude greater than SL during the rapid burning phase. During this rapid burning phase, since (rf – rb) is approximately constant, the flame front expansion speed and the mean burned gas expansion speed are essentially equal. The difference between uband Sb is the unburned gas speed ug just ahead of the flame front. Note that the ratio uf/Sbdecreases monotonically from a value equal to the expansion ratio eu/eb at spark to unity as the flame approaches the far wall.
Superimposed tracings of flame fronts. Illustration showing best fit circle to the 18th flame front.
Schematic diagram of combustion chamber geometry and spherical flame front. Angle versus distance, showing qualitative trajectories for flame center, flame fronts, and gas particle.
CONCLUSIONS Simultaneous pressure measurements and highs peed motion pictures of the visible flame in a spark ignition engine show that the initial flame front propagation speed is very close to that of a laminar flame for the same charge. As the flame grows, its speed increases rapidly to a quasi-steady value of order 10 times the laminar value. During the rapid quasi-steady propagation phase, a significant fraction of the gas entrained behind the visible flame front is unburned. The measurements also suggest that the final combustion phase can be approximated by an exponentially decreasing burning rate with a time constant of order 1 ms. Detailed analysis of the data has led to the development of a set of empirical differential equations that correlate well the experimental observations. The burning equations contain three parameters: the laminar burning speed of the charge SL, a characteristic speed uT, and a characteristic length lT. Measurements of SL under engine like conditions can be made in constant volume combustion bombs, and values for a number of common fuels are available. Values for uT and lT can be obtained from engine experiments, and preliminary correlation for relating these parameters to engine geometry and operating variables have been given. The data suggest that uT increases and lT decreases during compression of the unburned gas. For a given engine cycle, the parameters in the burning equations can be adjusted to fit the observed pressure curve. Cycle-to-cycle fluctuations in pressure can be caused by variations in any of the parameters SL, uT, and lT. Variations in SL can be caused by incomplete mixing of the fresh charge with burned residual gas in the cylinder and by variations in the stoichiometry of the fresh charge. Variations in uT, and lT are presumably associated with the statistical character of turbulence.
An additional parameter required to close the burning equations with a geometrical description of the enflamed region is the vector re giving the position of the apparent flame center. The nominal value of re is determined by spark plug position, but convection of the flame kernel at early times during propagation can produce significant displacement. It is observed that substantial cycle-to-cycle fluctuations can be caused by variations in the parameter re. Variations in re are presumably caused by convection of the initial flame kernel in the flow field near the spark plug. In this connection, it may be noted that a correlation between the pressure and the flow velocity .re near the spark has been observed in laser doppler measurements. Although the proposed empirical burning equations provide a relatively simple and accurate method of predicting the observed burning rates in spark ignition engines, the range of engine geometry and operating variables investigated in this experiment is relatively small and needs to be considerably extended. In particular, systematic investigations over a wide range of engine speeds, spark angles, valve lifts, and compression ratios are needed to establish the proper correlation for uT, and lT. the origin of the cyclic variations in the values of SL, uT , lT and re needs to be more closely examined, and correlation for relating their magnitudes to engine geometry and operating conditions need to be developed. The range of validity of the burning equations and their applicability, for example, to engines with significant swirl and squish also needs to be established. Finally, the experimental evidence presented and the proposed empirical equations need to be better understood in terms of the underlying physical mechanisms.
References • C. R. Ferguson, “Internal combustion engines applied thermodynamics” Wiley, 1985. • R. Stone, “Introduction to internal combustion engines” Macmillan Press, 1999. • J. B. Heywood, “Internal combustion engine fundamentals” McGraw Hill 1988. • G. P. Beretta, M. Rashidi, J. C. Keck, "Turbulent flame propagation and combustion in spark ignition engines". Combustion and flame, V52, N3, P217, 1983. • M. Rashidi, "The nature of cycle-by-cycle variation in the SI engine from high speed photographs". Combustion and flame. V42, P111, 1981. • M. Rashidi, "Calculation of equilibrium composition in combustion products." Journal of applied thermal engineering, V18, No 3-4, pp. 103-109, 1998 • M. Rashidi, "Measurement of flame velocity and entrained velocity from high speed photographs in the SI engine". The institution of mechanical engineers. Proceedings. V194, N21, P231, 1980. • M. Rashidi, "The sensitivity of elementary reactions for hydrogen oxygen in a well stirred rector". International journal of hydrogen energy. V5, P515, 1980 • M. Rashidi, M. S. Massoudi, "A study of the relationship of street level carbon monoxide concentrations to traffic parameters". Journal of atmospheric environment. V14, P27, 1980 • G. P. Beretta, M. Rashidi, J. C. Keck, "Thermodynamic analysis of turbulent combustion in a spark ignition engine; experimental evidence". Paper WSS/CI 80/20 presented at the spring meeting of the western states section of the combustion institute. April 1980 • M. Rashidi, M. S. Massoudi, "The use of gas engines for motor vehicles in oil exporting countries". SAE paper 770145, 1977.
