Given Geomentry:

Diagnosis:

Intersection Error:

While flagging cylinder head, we can find that there are intersection errors (indicated in pink). This is because the valves of

inlet side are superimposed on the head as it was lifted high. We need to translate it to the position where these intersection

errors are eliminated

Surface preparation and boundary flagging:

 

Inflow Boundary

Outflow Boundary

Piston 

 Liner

Intake and Exhaust Valves

First, the intake and exhaust valves are divided into three sections namely top, angle and bottom and are flagged

accordingly. Then the intake valve is translated down to eliminate intersection errors with cylinder head.

Exhaust Port

Intake port

Sark Plug and its terminal

Ring triangles translated

To remove the superimposing errors, we need to create boundary fences and extend the port boundaries close to the valve

boundaries. Make the three point arc on the circumference of the circle (Ring triangles – used to avoid deformation in valve

geometries during translation). This co-ordinates are taken and translated or pushed down by 1mm to remove the

intersection errors.

Cylinder head

Normal pointing

 

Completed Boundary flagging and regions:

Four regions are created:

Cylinder region:

Intake part 1:

Intake part 2:

exhaust region:

 

CASE SETUP FOR Full HYDRO SIMULATION:

IC Engine Application

 Parameters of PFI Engine

Materials

Gas Simulation

 

PARCEL SIMULATION

Enabling Parcel Simulation

In the parcel simulation, the predefined liquids option is clicked and the fuel IC8H18 is selected from the list of available

liquids.

Specifying species:

Gas species is taken care in reaction mechanism itself. Select parcel species as IC8H18

 Global Transport Parameters

Reaction Mechanism

(Imported from the given mech.dat file)

Run Parameters 

 Simulation Time Parameters

 Events:

The events are created between Intake , cylinderand exhaust region as given imported valve profile data.

The permanent event between intake port 1 and 2 are created.

Combustion equation of Isooctane :

C8H18 + 12.5 CO2 + 3.76 N2 8 CO2 + 9 H2O + (3.76*12.5) N2

Fuel Calculations :

RPM = 3000 rpm

Fuel mass/cycle = 3.00 E-05

RPS = 50

DPS = 18000

Time per Degree = 5.55556E-05 s

Time per 720 Degrees = 0.04 s

Fuel Flow rate = 7.5 E-04 kg/s

We assume that the combustion is stoichiometric and the combustion products mass fractions are calculated 

Stoichiometric Calculation

BOUNDARY CONDITIONS:

Inflow Boundary

 Intake Port – 1

Intake Port – 2

Intake Valve Top

Profile Configuration of intake valve

The profile configurations of intake and exhaust valves are given and this file is used. The min lift is kept as 2e-4 m and it is

cyclic.

Intake Valve Angle

Intake Valve Bottom

 Exhaust Port

Exhaust Valve Top

Profile Configuration of exhaust valve

Exhaust Valve Angle

 Exhaust Valve Bottom

Piston

Cylinder Head

 Liner

Spark Plug

 Spark Plug Terminal

Outflow

Physical Models:

 Turbulence Modelling

SPRAY MODELLING:

The spray model, combustion model and source/sink models are selected and it is attributed one by one.

From the given parameters, the fuel injection mass is to be found out and it should be given as input. Fuel sprays breaks up

into several parcels and we have to account those phenomena while defining the spray characteristics.

O’Rourke model in Turbulent Dispersion is for Gasoline and it used to analyse how turbulence affects droplet characteristics.

Use evaporation model is enabled because to capture the rate at which the parcel radius is going to change.

Maximum radius for ODE droplet heating is given as 1000. Here outside the sphere it is hotter and inside it is less hot due to

temperature distribution. So the sphere is discretised into cells and the 2D heat equation is applied to solve it. In case of

small sphere, entire sphere has same temperature and it is reasonable. But in case of bigger sphere, there will be

temperature gradient. So, outside the radius ODE is solved and inside it is not solved.

Urea model is not used because it is used for exhaust after treatment system which is beyond the scope of this simulation.

 

Nozzle Positions :

Nozzle diameter = 250 micro-meter

Circular injection radius = Nozzle radius

Spray cone angle = 10

• Nozzle 0

o center 0.0823357 0.00100001 0.07019

o Align Vector -0.732501 0.210489 -0.647408

• Nozzle 1

o center 0.0823357 -0.00099999 0.07019

o Align Vector -0.732501 -0.210489 -0.647408

• Nozzle 2

o center 0.0823357 -0.0004 0.07019

o Align Vector -0.5 -0.2 -0.647408

• Nozzle 3

o center 0.0823357 0.0003 0.07019

o Align Vector -0.5 0.2 -0.647408

 

Here, Isooctane (IC8H18) is added and its mass fraction is 1. The Rate shape values are given  to form the trapezoidal

profile. The peak injection pressure can be found out using Tools → Spray rate preview → Calculate rate.

Injector:

Rayleigh-Taylor model (RT) is used for solid cone sprays – gasoline and diesel.Discharge Co-efficient model is selected and

the value is given as 0.8.

The stability equation (or) dispersion equation gives us the time taken for the disturbance of the particular parcel to make it

unstable.

Model size constant is used to scale down original child droplet.

Then the Set recommendations is selected Gasoline PFI is chosen.

Injection Parameters :

1. Fuel flow rate = 7.5e-4 Kg/second

2. Injection start time = -480.0

3. Injection duration = 191.2

4. Fuel temperature = 330K

The start of injection and injection duration is given by the manufacturers. Injection takes place only at specific times and

single value of mass flow rate does not have sense and mass value should be absolute. For 4-stroke engine injection takes

place every 720 degrees. So for every 720 degrees, there is a particular amount of time for fuel to inject. That time is to be

taken.

Calculation :

RPM = 3000 rpm => RPS = 50 rps => Degrees Per Second (DPS) = 18000 degrees/s

Time per Degree = 5.55556E-05 s => Time per 720 Degrees = 0.04 s

Fuel Flow rate = 7.5 E-04 kg/s

Fuel mass/cycle = (Time per 720 degrees) * (Fuel Flow rate) = 3.00 E-05 kg

Therefore, the Total injected mass = 3 e-5 kg

 

Here, four nozzles are added from Nozzle 0 to Nozzle 3. For each Nozzle the configurations are done as shown in the above

figure. All the parameters listed above are given previously and it should be done for all the nozzles.

Nozzles are validated to check whether they are inside the computational domain. It is done in Spray Modelling window, Tools

→ Validate nozzle locations.

 

SOURCE/SINK MODELLING:

For setting up the spark, an energy source is introduced and it serves as the spark plug. By the energy equation, the energy

source increases the temperature of the species. And in the species transport equation, this energy is introduced. Thus, the

source term in the equation is responsible for the combustion to occur and its chemical kinetics affects the heat release,

products release for every cycle. Hence this concentrated energy source is to be introduced right at the compression stroke

for optimum combustion and the Source/Sink modelling is used.

Spark Ignition Parameters :

• Start of spark = -15 deg

• Spark duration = 10 deg

• Spark location = -0.003 0 0.0091

• Spark radius = 0.0005m

For source – 1, the energy value is given as 0.02 J and in Source – 2 also the energy value is 0.02 J during the start time (at

-15 degrees). Therefore the total is 0.04 J and it lasts for 0.5 degrees (upto -14.5 degrees) in a total of 10 degrees. Then the

energy from Source – 1 is shut off and the Source – 2 (0.02 J) is continued for the entire 10 degrees (upto -5 degrees).

 

 

COMBUSTION MODELLING:

Here, the SAGE model is used and the input parameters are shown in the above figure. The combustion start time should be

few degrees before spark starts (at -15 degrees spark starts). The combustion end time is determined by the degree at which

the exhaust valve opens. It is done by referring the exhaust valve profile.

The Multizone model tab near the SAGE Parameters tab is clicked and the Use multizone model option is checked for enabling

faster combustion.
 

 Grid Control

 Base Grid Size

 

Fixed Embedding Part 1:

 Intake Valve Angles

for the Intake Valves, the Boundary type embedding is used. The embedding is used to get the refinement near the intake

valve angles and the Boundary ID is given the same. The scale is 3 and the mesh size here will be 0.0005 m. It is fixed and

the same for exhaust valves also

Exhaust Valve Angles

 Big Cylinder Embedding

 

 Small Cylinder Embedding

To capture the combustion precisely, the combustion chamber (Cylinder – Boundary) is made with the Cylindrical embedding.

The small and big cylinders are used with the mentioned sizes with a scale of 1 and 2 with mesh sizes 0.002 m and 0.001 m

respectively

Cylinder embedding

Fixed Embedding part:2:

Small Spherical embedding – Spark

Large Spherical embedding – Spark

the Spherical embedding for precisely capturing the physics when the spark occurs. The small and large spheres are

embedded with the scale 3 and 4 and the mesh sizes will be 0.0005 m and 0.00025 m respectively for the maximum

coverage of gradient during the ignition and combustion process.

 

Injector embedding

For injection, the Injector type embedding is used. The Injector 0 is the Injector ID and the Radius 1, 2 and Length are given.

The scale is 4 so the mesh size will be 0.00025 m.

ADAPTIVE MESH REFINEMENT:

For the regions – Cylinder, Intake port – 1 and 2, the Velocity AMR is used. The start time, end time and Cycle period are as

mentioned in the above figure. The maximum cell count is 2000000 cells.

The Velocity AMR works as follows :

Max Embedding Level = 3

Converge uses Octry data structure and the formula for refinement:

Grid size = Base grid / 2embedded-level

50

Here the embedded level is 3 and the refinement is as follows,

Grid size = 0.004 / 23

= 0.004/ 8

Grid size = 0.0005m

Hence the smallest grid size in this simulation will be 0.0005 m

The Sub-grid criterion (SGS) used here is 1 m/s. Therefore for change in fluid velocity of 1 m/s in the mentioned regions the

mesh is refined.

For the regions – Cylinder, Intake port – 1 and 2, the Temperature AMR is also used. The start time, end time and Cycle

period are as mentioned in the above figure. The maximum cell count is 2000000 cells.

51

The Temperature AMR works as follows :

Here the embedded level is 3 and the refinement is as follows,

Grid size = 0.004 / 23

= 0.004/ 8

Grid size = 0.0005m

Hence the smallest grid size in this simulation will be 0.0005 m

The Sub-grid criterion (SGS) used here is 2.5 K. Therefore, for change in fluid temperature of 2.5 K in the mentioned regions

the mesh is refined.

Model coloured by regions

Finally, the case setup was finished and the input files are exported to the working directory. This complete PFI engine model

cannot be indulged into simulation using normal computers with 4 core – processors. It takes lot of computational time to

solve this model completely and this requires super computers like workstations with about 16 core – processors. Hence, the

extracted input files are given to the course instructor who has the previously mentioned supercomputer. The simulation was

run by them and the output files were shared. From that the 3D output files are generated using Converge Studio and the

results are post processed using the line plots from studio and 3D plots using ParaView. Results are shown in the following

section. By the end of this report, the link to the animations are provided. Several parameters like velocity, temperature,

pressure and other quantities are animated using ParaView.

Mesh:

X-normal

z-normal

The 3D animation of simulation:

 

The Temperature animation:

 

CONCLUSION:

In conclusion, Computational Fluid Dynamics Software – Converge Studio, was a useful method to simulate the fluid flow

behaviour with the relevant governing equation. The SI8 PFI Engine was successfully simulated using the given conditions

and specifications. The finer mesh size could provide the accurate results but, it increases the computational time

tremendously. So, tools like fixed embedding and adaptive mesh refinement were used to refine the mesh wherever and

whenever necessary. The physical models for capturing the combustion, sprays, sparks and turbulence are selected

appropriately for this case. To solve CFD cases like we did here, requires high computational power machines.

 

1. What is the compression ratio of this engine?

Generally, Compression ratio (r) of the engine is the ratio of total cylinder volume when the piston is at the bottom dead

centre (BDC) = VT, to the clearance volume = VC when piston at top dead centre (TDC).

r = VT / VC

r = (VC + VS) / VC , where VS is the swept volume.

The compression ratio of our engine is found using the volume plot show below i.e, by the ratio of maximum volume to the

minimum volume.

Volume Plot :

Maximum Volume = 0.00057 m3

Minimum Volume = 5.75 * 10-5 m3

Therefore, compression ratio (r) = Maximum Volume / Minimum Volume

= 0.00057 / 5.75 * 10-5

Compression ratio (r) = 9.9 : 1

2. Why do we need a wall heat transfer model? Why can't we predict the wall temperature from the CFD simulation?

Generally, the operation of an IC engine is based on chemical reactions taking place during the combustion. During power

stroke, the fuel-air mixture gets ignited (incase of spark ignition engine) by the spark plug and chemical reactions takes place

inside the combustion chamber. These reactions raise the temperature inside the chamber and forces the piston to expand.

During this process there will be an energy interaction between the fluid and the walls i.e, surfaces of the IC engine in the

form of heat. The heat transfer between the fluid and solid surface takes place generally in two modes – convection and

radiation. In general case CFD simulation, the Navier – Stokes equation is solved to get the results of the fluid flow. To get

accurate results, mesh size should be less and it increases computational time.

For this PFI engine simulation it takes lot of time to compute the case as it involves a complex geometry and processes. If we

want to capture the heat transfer – especially convection of heat from fluid to the solid surface, energy equations should also

be solved. This increases the time of simulation. For radiation of heat from the solid surface to atmosphere, an external mesh

is to be made to depict the external atmosphere and to solve the energy equation which increases the cell count. It makes

the simulation to run forever. Hence, we use conjugate heat transfer model where the solid part is solved in a steady state

using super-cycling option and the fluid part is solved in transient case.

For simulating complex geometries like IC engines, it takes lot of time for computing the fluid part itself in transient case

along with the combustion and reaction processes. So it becomes difficult to predict the wall temperature from the CFD

simulation as solving of heat transfer along with the Navier – Stokes equation consumes lot of computational power and

time. Wall heat transfer models are used to approximate this case. Basically, all the physical models used in a CFD simulation

uses some approximation to minimize computational time. Wall heat transfer models are used if we are focused only on heat

transfer giving less importance to fluid motions and reactions. For efficient operation of engine, the walls should be

maintained at a certain temperature. This is achieved by pumping coolant to take away the excess heat from cylinder wall.

3. Calculate the combustion efficiency of this engine.

Combustion efficiency is a measure of how effectively the heat content of a fuel is transferred into usable heat. In other

words combustion efficiency is the ratio of output energy derived from the engine to the input energy provided to the engine

during the combustion process.

Output Energy = 1241 J

Mass of fuel used per cycle = 3 * 10-5 kg

Lower Heat value of fuel (IC8H18) used = 44.8 MJ/kg

Potential energy in the fuel i.e, input energy = 44.8 * 10 6 * 3 * 10-5 = 1344 J

Combustion efficiency = Output energy / Input energy

= 1241 / 1344 = 0.9233

Combustion Efficiency of this engine = 92.33 %

4. Use the engine performance calculator, to determine the power and torque for this engine.

The engine performance calculator is used to analyse physical data and get useful results from the data. It is an useful tool to

calculate the quantities need to define an engine. In our case the calculator is opened and the file – thermo_region.out file is

added for the cylinder and then the engine.in file is loaded which contains the bore and stroke of the engine.

Duration of combustion = 240.199 degrees

Work done = 468.646 Nm

Indicated Mean Effective Pressure (IMEP) = 896428 Pa

CA 10(deg) = 6.83717 degrees

CA 50(deg) = 18.4623 degrees

CA 90(deg) = 31.701 degrees

Power is the rate of work done i.e, the amount of work done per unit time.

Power (P) = Work done / Tim

 

We know that, Time per Degree = 5.55e-5 s

Therefore, for one cycle of operation time taken (t) = 240.199 * 5.55e-5 = 0.013331 s

Power (P) = Work done / Time = 468.646 / 0.013331= 35154.48 J/s (or) 35154.48 W

Power (P) = 35.15 KW

Since Work and Torque has the same unit (N-m) , it doesn't mean value for Torque and Work will be the same. Both are

different quantities and has totally different meaning. So the value for work got from the engine performance calculator is not

the value for the torque. The Torque is calculated from the formula,

Power(P) = (2 * π * N * T) / 60

Torque (T) = (P * 60) / (2 * π * N)

The crank speed (N) = 3000 rpm – given at the beginning.

T = (35154.48 * 60) / (2 * π * 3000)

Torque (T) = 111.9 Nm

5. What is the significance of ca10, ca50 and ca90?

The values CA10, CA50, CA90, represents the Crank Angle at the completion of 10%, 50%, 90% of combustion. From these

values we can infer the time taken for the evaporation and combustion of the fuel. If the combustion process is fast,

whichrequires few crank angles and leads to increase in emissions. CA10 represents the start of ignition, CA50 represents the

end of fuel-air mixture combustion and at CA90 the burning of unburnt fuel is done by the propagating flame. If CA90 value

is large, the crank angles taken increases and the exhaust stroke tends to push the premature combustion products