Simulating the engine sector using converge

Introduction:

In the modern day cfd analysis engineer usually wants to do a lot of simulation for the design and calibration purposes. But simulation of the entire engine is not feasible since it more expensive computationally. In order to test the particular field of the engine domain for example looking for combustion modeling, emission control, flame interaction, spray penetration etc. For this engine sector simulation is performed. Here we take the engine combustion chamber such as cylinder head, liner and piston, and cut in the pie angle in order to save the computational time. The angle which it is cut depends on the no. of nozzles in the injector if the injector has 6 nozzle than the sector will be 60 degree i.e. 360/6, if it has 8 nozzle than 45 so on and so forth.

Here the challenge is to simulate the diesel fuel with 2 piston design i.e.

1. Open – W piston
2. Omega (reentering) piston

In converge there is the feature we can call extract profile from Surface.dat as shown in figure below

Different bowl profile is shown below

1. Open W
2. Omega

Engine Parameter

• Cylinder bore : 0.13716 m
• Stroke : 0.1651 m
• Con rod length : 0.263 m
• Rpm : 1600
• Fuel: diesel

Boundary Conditions

• Piston Temperature : 553 K
• Cylinder Wall : 433 K
• Cylinder Head: 523K

Here the different number of nozzles were performed

1. Omega Piston : 8 nozzles i.e. using 45 degree sector mesh
2. Open W : 6 nozzles i.e. using 60 degree sector mesh.

The reason for this is that want to check the effects of increasing the no of nozzle on the spray penetration with also fuel atomization.

Nozzle parameters:

For the simplicity I am considering the parameters of the nozzle such as position cone angle. As shown in figure below

For simplicity the spray injection is also kept same for both cases

Fuel Injection Parameters

1. Injection start: - 9.0 degrees
2. Injection duration: 21 degrees
3. Total Injected Mass: 2.70167x10^-05 Kg
4. Fuel Temperature: 341 K

Here is the snap shot for the spray pattern

Model used for capturing spray parcel penetration and emission are shown in the following snap shot, the parameters were kept same since

Converge recommends for SAGE for IC Engine combustion modelling so in our case we will also use the same The following figure describes the SAGE configuration

For Emission since this the diesel more focus is on the SOOT and NOX emission for this reason we will use the Hiroyasu soot model and zeldovich mechanism for NOx emission. These are most extremely popular formation mechanism for emission, if you want more information Hiroyasu check the link for more information at the end of this report.

Turbulence modeling was setup up to same RNG k and epsilon with all the parameter was kept as default i.e. recommended by converge. Following is the snap shot

Meshing setup:

Since it is the sector mesh we can afford to have the smaller mesh since this kind of configuration is not computational heavy so for base grid setup the cell was kept :- 0.0014 m in all the directions since the NOx and SOOT are more sensitive it’s better to setup the AMR for temperature and velocity respectively

This helps to eliminate the error for temperature which is generated from the source terms of the convections and diffusions of the fuel mechanisms.

Results

Mesh:

Following is the mesh for Omega piston on left and Open W on the right. Here is the snap shot

Since it is important to check the timing of the simulation for making sure there is no mistake to here is the snap shot so that we know which sequence will be next.

Since we are doing the engine simulation we might check the performance and how much fuel is unburnt

1. So lets check the Peak pressure of different piston configuration since we are dealing with the diesel engine which uses auto ignition it is very important to do check the parameter

So as we can check from the figure for the omega piston require more pressure than the Open W piston i.e. means that fuel is getting atomize more efficiently on the omega piston configuration rather than former one. So lets check the temperature and emissions such as SOOT and NOX

Temperature:

From the above figure we can see that fuel is burnt properly in the initial phase in Omega rather than Open_W.

Integrated HRR:

Difference between the integrated HRR as shown below

Open W (J)	1.1704577E+04
Omega (J)	9.6557352E+03
Difference (J)	2.0488418E+03

As you can see the Open W has more Integrated HRR and difference is also big so if we consider the fuel injected each cycle i.e. 2.70167x10^-05 Kg and multiplied with the Qlhv of the fuel in this case I am considering the 43.2 MJ/Kg.

Calculation

And for Integrated HRR will be 2048.841 KJ

For 1000 cycle there will be how much fuel left un burnt will be 0.0474268 Kg which is very small compare to the fuel which we are adding. This can be confirm by looking at the diesel conversion graph in gas phase

Following video of Spray and Temperature will show us the better view respectively.

1. Spray Mesh
2. Temperature

Emissions since NOx is extra sensitive to Temperature I suspect that the NOx will be Highre in Omega piston compare to Open W

NOx Graph

NOx Formation Animation

SOOT Graph

SOOT formation animation

Performance Calculation

Omega Piston	Open_W

As shown in the above performance calculation we can see that performance wise the Omega has better performance parameter i.e. Work Done, IMEP etc. so there is trade of between if we are more focused on the performance than omega but if we are more concern about the emission than definitely Open W piston configuration is the best but this cannot be said firmly since it will require more simulation for data accuracy and compare it to the experimental setup. If we want to improve we can one of the ways is getting the nozzle angle change for the OPEN W piston and see if it’s interfering the results. Or we change introduce the newer mechanisms for the emission.

Here is the video of the Diesel evaporation for both the pistons

Literature Links

https://link.springer.com/article/10.1007%2Fs12046-018-0907-x

https://www.jstor.org/stable/44644056