Emission characterization on a CAT3410 engine

In this project, omega and openw piston diesel engine are simulated using sector approach. Sector based simulations can be applied to closed cycle piston, where the intake and the exhaust ports are neglected to simplify geometry for computational efficiency. Here, a part of the cylinder region is taken ( a sector ) and simulated. First the bowl profiles of the engine are extracted (using the extract option in make engine sector surface tool) and the sector geometries are then created using the make engine sector surface tool with a compression ratio of 17.5, cylinder bore - 0.13716m, cylinder stroke - 0.1651m and connecting rod length - 0.263m. The sector taken for the simulation here has an angle of 60degree between the front and the back face. Also called the sector angle. For both piston types same settings were used to allow for comparison.

The geometry of omega piston:

The geometry of openw piston:

The mesh of the omega piston:

The mesh of the openw piston:

The following boundaries are created (the boundaries are highlightened for identification):

Piston:

For omega:

For openw:

Cylinder head:

For omega and openw (looks the same so single image is provided):

Cylinder wall (cylinder liner):

For omega and openw (looks the same so single image is provided):

Front face:

For omega:

For openw:

Back face:

For omega:

For openw:

Application is set as IC engine and the engine parameters were taken as follow:

cylinder bore - 0.13716m

cylinder stroke - 0.1651m

connecting rod length - 0.263m

speed - 1600rpm

swirl ratio - 0.978

swirl profile - 3.11

Parcel simulation is enabled with predefined liquid DIESEL2 and appropriate reaction mechanisms were used which are given by Skill-Lync.

In the species, DIESEL2 is included under parcel and HIROY_SOOT and NOX were included under passive with a Schmidt number of 0.78. The simulation time parameters are taken as below:

Start time : -147.0deg

End time : 135.0deg

Initial time-step : 5e-07s

Maximum time-step : 2.5e-05s

Minimum time-step : 1e-08s

Boundary conditions:

Piston : Assigned as a moving wall boundary following piston motion with a temperature of 553k. Law of wall boundary condition is used here.

Piston motion option moves the selected boundary (The wall should be set as moving) with the velocity and direction of the piston motion. The piston motion is generated internally by CONVERGE, based on the geometry and engine parameters.

Front face : Periodic boundary with sector shape and an angle of 60deg.

A PERIODIC boundary type runs the simulation for one portion of the domain and replicates the results obtained to other identical portions of the domain by doing so it reduces computational time.

Back face : This boundary is matched with front face boundary. (Matching boundary)

All of the Boundary Conditions for the Matching Boundary will be the same as those specified with boundary to which it was matched. So, back face will have same boundary conditions as front face.

Cylinder wall : Cylinder wall is a stationary wall boundary following law of wall boundary condition with a temperature of 433k.

Cylinder head : Cylinder head is also a stationary wall boundary following law of wall boundary condition with a temperature of 523k.

Regions:

Since this is a closed cycle simulation only one region is required.

Incylinder region: This region is initialised with O2, N2, CO2 and H2O making up air of mass fraction 1 as follows:

O2 - 0.23029

N2 - 0.76765

H2O - 0.0006296

CO2 - 0.0014304

Passive HIROY_SOOT and NOX are also initialised with a mass fraction of 0.

The initial temperature of the region is taken as 355k with a pressure of 1.97bar. The turbulent kinetic energy of this region is taken as 62.0271`m^2/s^2` and turbulent dissipation is taken as 17183.4`m^2/s^3`.

Spray is modelled such that the parcels cluster near spray cone centre. For turbulent dispersion O'Rourke model is adopted. Frossling evaporation model is used with evaporation sourced from C7H16. Radial temperature distribution is set to activate at 1000m droplet radius i.e., essentially uniform temperature distribution is used for the simulation. No urea after treatment models are considered. For measuring the liquid spray penetration a mass fraction of 0.97 is taken with the minimum mass fraction required to activate vapour penetration taken as 0.001 and a bin size of 0.001m. Mass diffusivity constants for DIESEL2 (kerosene defaults) were considered for the simulation.

NTC collision with post collision outcome model is selected with dynamic drop drag model having default constants. Breakup model is included for the simulation. The wall interaction is set to rebound/slide. From the given parameters the injection pressure automatically calculated by Converge is found to be 1097.05bar.

The y-axis here is the proportional velocity whereas the x-axis represents normalised time.

Kelvin-Helmholtz model and Rayleigh-Taylor model are selected with recommended values for diesel and discharge coefficient is taken as 0.7. The injection is set for a cyclic period of 720deg. The injection parameters are taken as follow:

1. Total mass injected per cycle = 2.70167e-05kg
2. Injection start time = -9.0deg
3. Injection duration = 21.0deg
4. Fuel temperature = 341.0k

and the following are Nozzle parameters:

1. Nozzle diameter : 0.000259m
2. Circular injection radius : 0.0001295m
3. Spray cone angle : 9deg
4. Nozzle radial position : 0.00097m
5. Tilt angle in the xz plane : 62.5deg

After this, SAGE combustion modelling is enabled with the following parameters:

1. Start time(crank angle):-10.0deg
2. End time(crank angle):135.0deg
3. Minimum cell temperature(for SAGE to activate):600.0k
4. Minimum HC mass fraction in cell(for SAGE to activate):1e-08

Multizone model is used with 2 dimensions temperature and equivalence ratio (react. ratio) with bin sizes 5k and 0.05 respectively.RANS based RNG k-epsilon model is used for turbulence modelling with default values.

Mesh is generated with a base grid size of 0.0014m, Adaptive mesh refinement is enables for velocity and temperature with a max cell count of 2000000 in the incylinder region. For velocity amr is enabled permanently with max embedding level of 2 and sub grid criterion of also 2. Temperature amr is enabled sequentially with a max embedding level of 2 and sub grid criterion of 5 starting from -12.0deg till 180.0deg. (Since only part of the cycle is run, temporal type of sequential and cyclic will have the same effect).

Fixed embeddings are also provided to see the amr into activating.

Fixed embedding 1:

Nozzle is embedded with NOZZLE embedded type with 0.001m radius of the first end of embedding( start of the spray) and 0.003m radius of the second end. Taking 0.01m as truncated cone embedding length. The embed scale is taken as 2 with sequential temporal type having start time as -12deg and end time as 15deg.

Fixed embedding 2 and 3:

Boundary embedding is applied for piston and the cylinder head with a embed scale of 1 and number of layers 1. Again sequential temporal type is used with a start time of -20deg and a end time of 180deg.

Full hydrodynamic simualtion of the above case setup is simulated and the following results are obtained:

NOX emission graph showing comparison between omega and openw piston:

From the above graph we can see that NOx levels poduced by the openw engine are very low compared to omega piston engine, looking into it further we find that from the mean temperature graph omega piston seems to have reached much higher temperature, this implies that at high temperature more NOx is produced.

Mean temperature graph showing comparison between omega and openw piston:

Openw did not reach higher temperature, this could be because heat release rate is not as high as in omega piston engine case which becomes evident after observing the graph attached below.

Heat release rate graph showing comparison between omega and openw piston:

Although from the integrated heat release rate at the end (seen from the graph below) seems to be the same, from -5deg to 30deg (time when combustion is taking place) in the heat release rate graph the high heat is released in the omega piston which leads to high temperature and high NOx formation.

Integrated heat release rate graph showing comparison between omega and openw piston:

NOX emission animated in a slice of sector to compare between omega and openw piston:

Fuel injection animated in a slice of sector to compare between omega and openw piston:

Fuel(C7H16) graph showing comparison between omega and openw piston:

Soot emission animated in a slice of sector to compare between omega and openw piston:

Soot graph showing comparison between omega and openw piston:

From observing the fuel cutplane (slice) of sector animation, as we had initialised wall interaction as rebound/slide, the fuel in case of the openw piston impinges and slides on the surface of the piston. But, compared to the omega piston, the spray in case of the openw piston has little time to vaporize as it impinges shorty after getting injected as compared to the omega piston. This leads to reduced evaporation in the cylinder region thereby leading to low heat release rate initially thereby contributing to low temperature in the region which results in low NOx. This type of piston with the same nozzle angle can be used if there is a need to reduce NOx pollution but this leads to incomplete combustion which results in high soot production which can be seen from the above graph. So, reducing NOx pollution and reducing soot emission is a balancing act.

But if the objective is to increase efficiency of combustion, the nozzle angle (Tilt angle in the xz plane) can be increased so that the spray impinges the piston surface at a much farther distance resulting in more time for evaporation of the fuel spray.

Other emissions:

CO:

CO graph showing comparison between omega and openw piston:

CO emission animated in a slice of sector to compare between omega and openw piston:

HC:

HC emission (unburnt hydrocarbon) comparison between omega and openw piston:

Other parameters:

Liquid spray mass comparison between omega and openw piston:

From the engine performance calculator the following values were calculated:

Omega piston:

Work : 3409.28Nm IMEP : 1.39675e+06Pa

Openw piston:

Work : 3028.53Nm IMEP : 1.24076e+06Pa

Pressure graph showing comparison between omega and openw piston engine: