Quantitative comparisons are made against the experimental data [22]。 In order to make comparisons

ENGINE SIMULATION

The engine used for the simulation is a single-cylinder version of the Caterpillar 3400 series diesel engine operating on the baseline mode。 The engine specifications are shown in Table 1。

Figure 4。 Temperature versus radial location at several downstream locations

Shown in Figure 5 are the estimated root mean square (RMS) temperature, Trms, as determined from Equation (42):

Cylinder bore X stroke (mm) 137。6 X 165。1 Connecting rod length (mm) 261。62

Displacement volume (L) 2。44

Compression ratio 15。1

Number of nozzle X diameter (mm) 6 X 0。259 Spray angle 27。5o

Intake valve closure -147o ATDC

Injection Pressure 90 MPa

Engine speed (rpm) 1600

% Load 75

Injected mass(g/min) 129

Start of injection -9o ATDC

Duration of injection 21o

Table 1。 Caterpillar 3400 series engine specifications

where < > represents an averaging operator。 For example, <T> means that resolved temperatures,  T˜ , are averaged in time and then in azimuthal direction to obtain an ensemble average。 Experimental data is also shown for comparison。 It should be noted that computed results determined by Equation (42) are conceptually different from the experimental data。 Unlike experimental data, which deals with the fluctuations of instantaneous values, the computed data indicates fluctuating component of LES resolved value。 It can be seen that the computed Trms and experimental Trms have similar shape although the discrepancy in terms of magnitude tends to be larger at downstream locations (z/D > 30)。

A one sixth sector mesh (~30,000 cell at IVC), shown in Figure 6, is used for computational efficiency。 At TDC, the numbers of cells in radial, azimuthal, and axial directions are 19, 30, and 10 respectively。 This corresponds to the grid resolution of 2。6 mm in radial, 2o in azimuthal and 1 mm in axial direction respectively。 This mesh might be relatively coarse for a LES application。 However, this mesh is used not only due to computational efficiency but also, by its nature, LES responds to the grid size through the filtering length scale。 Nevertheless, in future LES engine simulations, a finer truly 3-dimensional mesh should be used for improvements。

Figure 6。 Perspective view of computational mesh with spray droplets near the top dead center

Figure 5。 Temperature RMS versus axial distance from jet exit

The engine simulation was started at  intake valve closure and ends at 50o ATDC。 The  in-cylinder conditions at this time, including  temperatures, pressures  and  residual  species,  are  determined  by a

one dimensional gas dynamics code [24]。 The initial conditions for both subgrid kinetic energy and subgrid mixture fraction are set to negligibly small values。 It should be noted that multiple-cycle simulations are preferred in a more detailed study in order to remove the effects of these initial conditions and generate realistic turbulent structures。 The pressure trace and the heat release rate are compared to experimental data in Figure 7。 There is an excellent agreement between LES simulation results and experimental values  indicating that the models capture the overall behavior correctly。