Week 10: Project 1 - FULL HYDRO case set up (PFI)
Introduction:
In a spark-ignition engine, the fuel and air are mixed in the intake port. Then it goes to the cylinder through the intake valve. Inside the cylinder, mixing with residual gas takes place and then compressed. After the end of the compression stroke, combustion is initiated through the spark plug which produces electrical discharge. Following inflammation, a turbulent flame develops and propagates through the premixed fuel until it reaches the combustion chamber walls. Then the burned gas is removed by exhaust stoke.
Spray modeling:
Converge can model both gaseous and liquid sprays. For Gaseous spray, Converge uses Eulerian solver. For the liquid sprays, Converge uses the Lagrangian model to the discrete parcels and Eulerian model to the continuous flow domain.
The Lagrangian specification of the flow field is a way of looking at fluid motion where the observer follows an individual fluid parcel as it moves through space and time.
The Eulerian specification of the flow field is a way of looking at the fluid motion that focuses on specific locations in the space through which the fluid flows as time passes.
Parcels:
the fluid parcel is a very small amount of fluid, identifiable throughout its dynamic history while moving with the fluid flow. Each Parcel represents a group of identical drops that have the same radius, velocity, temperature, etc. Converge solves every parameter on per parcel parameter rather than per drop basis. Because the drops in a parcel are identical.
SAGE Combustion Solver:
It is the detailed chemistry solver that uses local conditions to calculate reaction rates based on the principle of chemical kinetics. Using this solver, the chemical reaction can be performed in every cell. The elementary reaction in the mech.dat is in Arrhenius format.
Multizone model:
The multizone model reduces the number of SAGE calculations by grouping the cells into bins. Using this method, the calculation can be done for bins instead of solving every cell.
Objective:
To simulate the Port Fuel Injection gasoline engine combustion process with given Boundary condition and initial condition using CONVERGE.
The geometry of the Cylinder:
- Cylinder bore = 0.086 m
- Stroke = 0.09 m
- Connecting rod length = 0.18 m
- Crank speed = 3000 rpm
Case setup:
1) Materials:
- Gas simulation is also added.
- Parcel simulation is also added.
- Parcel - Iso-Octane (IC8H18)
- Iso-Octane has the physical property which is very close to the gasoline.
- Reaction mechanism - mech.dat file (it contains reaction mechanism for gasoline)
- Thermodynamic data - them.dat (it contains thermodynamic data of species which present in mech.dat file)
2) Simulation parameters:
- Simulation mode - Hydrodynamic
- Start time = -520 degree
- End time = 120 degree
- Initial time step = 1e-7 s
- Minimum time step = 1e-8
- Maximum time step = 0.0001 s
3) Boundary condition:
Marked Boundary conditions
4) Initial condition end Events:
The intake port is divided into two regions. A region that is near to the cylinder is marked as Intake port 1. A region that is near to the inlet is marked as Intake Port 2. The reason for splitting the intake port into two regions is initially Intake port 1 should have some amount of air-fuel mixture. To assign initial conditions separately, two separate regions of intake port are necessary.
Note:
Navier-Stokes equations are the governing equations of fluid dynamics and in general, can be solved only numerically. In this method, the computational domain is being decomposed at a number of small finite volumes or computational cells. This process of decomposition is called meshing. The bodies in the flow are often treated as voids. The surface of the body becomes one of the boundaries of the computational domain and the appropriate boundary conditions must be set on it to obtain the numerical solution. If the flow is compressible, the momentum equation needs to be solved. In order to solve the momentum equation, certain boundary conditions to be assigned like a slip or no-slip boundary conditions. If the flow is incompressible, the additional energy equation is also needed. In order to solve the energy equation, boundary conditions are necessary.
If the solid walls of the cylinder are considered, applying predicted temperature value is not necessary. Using Conjugated heat transfer model, it can be able to solve without predicted temperature value. But in our case, predicted temperature value for all the walls is needed.
Events:
Events are used to create the virtual open and close diaphragm between cylinder and Inlet/Exhaust Ports. Disconnect triangles will be created when the event is marked as close at the gap between the valve top and the port sheet region.
5) Physical models:
Model 1: Turbulence model
RNG k-epsilon model
Model 2: Spray model
- "Distribute the parcel evenly throughout the cone" option is chosen for parcel distribution. It produces evenly distributed parcels throughout the cone rather than clustered parcels.
- O' Rourke Model shows the best result for gasoline engine simulations.
- The evaporation model mainly governs the rate of change in the parcel radius due to evaporation effects.
- "Source all the base species" option is chosen as the evaporation source. It converts the liquid Iso-Octane into gaseous Iso-Octane when evaporation takes place.
- No-time-counter (NTC) method can be used for the general case of varying numbers of droplets per parcel. The NTC method is several orders of magnitude faster and slightly more accurate than O'Rourke's method for several test cases.
Injectors:
A fuel injector is placed in the location of the Inlet port 1. It injects the fuel into the inlet port region.
Injector rate shape profile setting:
| Crank angle | rate shape value |
| 0 | 0 |
| 1 | 1 |
| 2 | 1 |
| 3 | 0 |
Converge uses this above profile to inject fuel in a given amount of time.
Model:
- Kelvin-Helmholtz (KH) model
- Rayleigh-Taylor (RT) model
The KH-RT hybrid atomization model consists of two distinct steps: primary and secondary breakup. The Kelvin-Helmholtz (KH) instability model was used to predict the primary breakup of the intact liquid core of a fuel jet. The secondary breakup of the individual drops was modeled with the Kelvin-Helmholtz model in conjunction with the Rayleigh-Taylor (RT) accelerative instability mode. A modification was made to the KH-RT hybrid model that allowed the RT accelerative instabilities to affect all drops outside the intact liquid core of the jet.
[ Total Injected mass in Kg/Cycle calculation:
Fuel flow rate = 7.5e-4 Kg/s
RPM = 3000
RPS = RPM/60 = 3000/60 = 50
DPS = RPS*360 = 50*360 = 18000 deg/s
Time per degree = `DPS^-1` = 5.555e-5 s/deg
Time per 720 degrees ( for 1 cycle) = 5.555e-5 * 720 = 0.04
Fuel mass per cycle = 7.5e-4 * 0.04 = 3e-5 Kg/cycle]
From the previous data, Converge calculates peak injection pressure and velocity.
Model 3: Combustion model
- CONVERGE’s SAGE detailed chemistry solver uses local conditions to calculate reaction rates based on the principles of chemical kinetics. This solver is fully coupled to the flow solver, but the chemistry and flow solvers parallelize independently of one another, which speeds up the simulation.
- Detailed chemistry modeling can be expensive, and so CONVERGE incorporates acceleration strategies such as the Multi-zone model.
Model 4: Source/Sink model
- Two energy sources are added to the source model in the shape of the sphere.
- It mainly replicates the spark plug. It produces heat energy in the cyclic period. As a result of these energy sources, the temperature is increasing. And then it affects the chemical kinetics which will produce child species based on the local condition.
6) Mesh Parameters:
Base mesh:
- dx = 0.004 m
- dy = 0.004 m
- dz = 0.004 m
Fixed embedding:
- Fixed embedding is the tool to make the mesh finer at the given Boundary/Region.
- Fixed embedding for Intake /Exhaust valve angle is to provide minimum lift while valves are moving.
- To provide fine mesh inside the cylinder region, fixed embedding is given to the cylinder region with a scale of 2. It provides a 1 mm grid size inside the cylinder.
- The fixed embedding is enabled for the sparkplug region to calculate the temperature gradient accurately in that region.
- Small spherical embedding provides the fine grid with a scale of 5 closely near to the spark. Large spherical embedding provides the fine grid with a scale of 5.
- Converge provides exclusive fixed embedding for injectors. The conical-shaped region was marked to provide fine mesh with a scale of 4.
Adaptive Mesh Refinement:
- Adaptive mesh refinement is used to get the fine mesh at the desired location based on the flow parameters.
- It monitors the curvature (`(del(psi^2))/(del(x^2)` of the mentioned flow parameter when it exceeds the given value then it produce fine mesh on the particular cell.
- For this simulation, we gave AMR for velocity and temperature.
Procedure:
1) The above condition has been given to Converge Studio to create input data files.
2) Input data files are accessed by using the Cygwin terminal.
3) The simulation has been performed using Converge-3.0 executable. To decrease the simulation time, 2 cores are used with the help of MSMPI.
4) The output files are converted into Paraview VTK in-line binary format by using Converge Studio Post Processing 3D to see the results in Paraview.
Results:
1) Volume vs Crank Angle plot:
- Compression ratio = Total volume/Clearance volume
- The compression ratio (CR) is defined as the ratio of the volume of the cylinder and its clearance volume.
- Total volume = Maximum volume = 5.7551216e-5 m^3
- Clearance volume = Minimum volume = 5.7442919e-4 m^3`(2*pi*N*T)/60`
- Compression ratio = 10.02
2) Combustion efficiency of the engine:
Combustion efficiency is the ratio of total heat released (Integrated heat released) to the total heat added ( Calorific value of the fuel).
Combustion efficiency = (Total heat released per cycle/Total heat added per cycle)*100 %
- The calorific value of the fuel (Iso-Octane) = 47.77 MJ/ kg
- Amount of fuel added per cycle = 3e-5 kg
- Total heat added per cycle = Calorific value of the fuel*Amount of fuel added per cycle = 3e-5*47.77e6 J/kg = 1433.3 J
- Total heat released per cycle = 1241.14 J
Combustion efficiency = (1241.14/1433.3)*100 = 86.59 %
3) Power and Torque calculation:
Engine Performance data:
Mechanical power can be defined as the time derivative of work. It says how much work has been done in a particular amount of time. Torque is a rotating force produced by an engine's crankshaft
Power = Work/Time
- Gross work from engine performance calculator = 468.646 J
- Time per degree of rotation = 5.555e-5 s/deg
- Time per 240.2 degrees of rotation = 5.555e-5 * 240.2 = 1334.311 e-5 s = 0.013343 s
Power = 468.646/0.013343 = 35123 J/s = 35.123 kW
Power = `(2*pi*N*T)/60`
Torque = `(P*60)/(2*pi*N)`
- N = 3000 rpm
- Power = 35.123 kW
Torque = `(35123*60)/(2*pi*3000)` = 111.8 Nm
4) CA10, CA50, and CA90:
The CA10, CA50, and CA90 are defined as the position of the crank angle at which the cumulative heat release reaches 10%, 50%, and 90% respectively. The physical significance of CA10 is
that it is used to identify the start of ignition. whereas the CA50 indicates the end of premixed combustion and the start of diffusion combustion of fuel used in the engine. Higher the difference between CA10 and CA50 (CA50-CA10), higher will be the combustion duration and vice versa.
5) In-cylinder pressure:
Inference:
- The pressure inside the cylinder reaches the maximum pressure of 3.887 MPa at 23.5 degrees of crank angle.
6) In-cylinder temperature:
Inference:
- The temperature inside the cylinder reaches the maximum pressure of 2481.11 K at 32.63 degrees of crank angle.
7) Heat released rate (J/CA)
Inference:
- It clearly shows how much heat is released at each and every crank angle. It reached maximum value after the combustion process starts.
8) Integrated Heat Released (J):
Inference:
- It represents the time integral of heat release. Here time means crank angle. After the end of the cycle, 1241.14 J amount of heat released.
9) Mass of the fuel vs Crank Angle
Inference:
- It shows the mass of the liquid fuel parcels which are injected inside the intake port 1 region through the 4 nozzles.
10) Carbon monoxide emission:
Inference:
- It shows the carbon monoxide emission. After combustion starts, it reaches the maximum amount. But slowly those CO molecules react with other combustion products. At the end of the combustion, It produces around 7e-6 kg of CO per cycle.
11) Carbon dioxide emission:
Inference:
- It shows the carbon dioxide plot. As we initiated the simulation with a certain amount of combustion products inside the cylinder, it shows the CO2 at the begging of the cycle.
- At the end of the cycle, 8.50504e-5 kg of CO2 is produced as a result of combustion.
12) Hydrocarbon vs Crank angle:
Inference:
- It shows the mass of the hydrocarbon vs Crank angle.
13) Hiroy Soot:
Inference:
- Soot is a mass of impure carbon particles resulting from the incomplete combustion of hydrocarbons. At the end of the cycle, 2.5e-8 kg of soot is produced.
14) NOx emission:
Inference:
- It shows the NOx emission vs Crank angle plot. NOx is one of the combustion products. At the end of the cycle, 1.3e-6 kg of NOx is produced.
15) P-V Diagram:
Inference:
- The above diagram shows the pressure vs volume plot. It is one of the famous plots to see the entire combustion process. In thermodynamics, it shows the full complete cycle of the combustion process. It starts and ends at the same point. Our obtained PV diagram approximately matches with the ideal PV diagram.
16) Mesh animation:
Inference:
- From the above animation, it clearly shows mesh refinement due to AMR and Fixed embedding in the cylinder region.
17) Parcel Animation:
Inference:
- It clearly shows the fuel parcel injected from the fuel injector to the inside the cylinder.
18) Temperature contour animation:
Inference:
- It shows the temperature contour of all-region. from the above animation, it also represents how the flame is propagating through the cylinder.
Conclusion:
Simulation of a Port fuel injection gasoline engine has been completed successfully. The results of this simulation approximately meet our basic expectations as expected.