OCR Text |
Show Each packet is described by an entry index, I, and a packet classification index, J. The first packets injected into the cylinder have an entry index of 1; the second have an entry index of 2; and so forth. At each entry time, the packets are also identified with a classification index (e.g., 1, 2, ...). The energy or combustion term, 0Comb' *s composed of three constituents which represent the energy release from the coal, volatiles and liquid fuel vapor combustion. During injection, a charge of fuel mixture which is composed of specified amounts of liquid carrier and coal enters the cylinder. The coal consists of a specified mass fraction of char and volatiles. The particle combustion model is similar to the one developed by Caton and Rosegay (1983) and modified to include devolatilization by Bell and Caton (1984). This model assumes spherical particles with uniform but time varying temperatures. Also, this model includes variable particle heating values and variable particle densities due to devolatilization. Figure 2 shows schematically the coal particle combustion process which includes solid phase reactions, gas phase reactions, and devolatilization. The particles are assumed to burn via two mechanisms yielding carbon monoxide, C(5) + » 02(g) * "(g) <5> and C(s) + C02(9) * 2C0(9) (6) The carbon monoxide is then assumed to immediately mix with available oxygen in the packet and burn in the gas phase away from the particle. By simultaneously satisfying the mass conservation equations and surface kinetics of the particle, the reaction rates are calculated (Caton and Rosegay 1983b; Rosegay and Caton, 1983). A second constituent to the combustion term in the energy equation is from volatile combustion. A detailed description of the devolatilization model may be found elsewhere (Bell and Caton, 1984; Bell and Caton, 1985a) and, therefore, is only summarized here. Based on the assumptions that the volatiles are locally released and "jet" away from the particle, a simple model was formulated for the devolatilization process. Once evolved, the volatiles are assumed to react with available oxygen in the packet (after an ignition delay) far from the particle. The assumption that the volatiles are locally released from the particle has been supported by several experimental observations (e.g., Matthews and Street, 1984; Seeker et al., 1981; McLean et al., 1980). If thermo-chemical conditions are sufficient, simultaneous solid reaction and devolatilization may occur. Since fundamental, mechanistic rate expressions for devolatilization are not available due to the complexities of the process, empirical data were used which have been approximated with first order reactions such as: ft' Vl + °Zk2> F <Vn«x " V» <7> c where the rate constants, kj, are given by Arrhenuis expressions. For the current study, the rate constants recommended by Kobayashi et al. (1977) were used and 50% of the coal mass was assumed to be the maximum volatile yield. A final constituent of the combustion term is the liquid fuel combustion. For simplicity, the steady-state, isolated, single drop approach of Spalding (1955) was used in the model. The vaporization submodel is described in detail elsewhere (Bell and Caton, 1985a). The final result gives a relation for the mass flux leaving the surface of the drop. The vaporized fuel then mixes with available oxygen in the packet and reacts after an ignition delay. The ignition delay correlation is used to determine the time between when the fuel is injected and the vaporized mixture ignites. This time is characterized by a slight gas cooling due to the evaporation. (The correlation used here for the fuel vapor and volatiles should not be confused with the overall engine ignition delay times which are based on the net effect of all the fuel on pressure recovery in the cylinder after injection.) The ignition delay correlation used in this study was that of Spadaccini and TeVelde (1982) and is of the form: T • -\ exp (-£-) (8) Pn RT where, A = 2.43 x IO- 9 , E » 41.56 kcal/mole, and n»2. This correlation has been used sucessfully in other cycle simulations (Mansouri et al., 1982). Other terras in Equation (3) such as piston work were calculated with a knowledge of the instantaneous cylinder conditions. The convective heat transfer includes convection from the particle and heat losses to the droplet and cylinder walls. The heat losses to the walls were calculated using an empirical correlation from Woschni (1967) based on the average cylinder gas temperature. Heat transfer between the packets and the unburned zone was based on the volume (surface area) and temperature of the specific packet. Depending on the temperature of the packet this heat transfer may be to or from the packet. In addition to heat transfer across packet boundries, packet work due to expansion or cooling is also transferred between packets thus allowing some communication between packets. Droplet and particle convection rates were based on coefficients corresponding to particles and droplets in a quiescent atmosphere. Radiation losses from the particle to the cylinder wall were also included in the model. RESULTS As previously discussed, the results of this study are for particular engine and fuel specifications. The engine and operating parameters selected for this study were chosen to correspond to locomotive engine designs. Table 1 lists the engine . |