Port Fuel Injected Spark-Ignited Engine Simulation
Objective
To simulate and analyze a Port Fuel Injected Spark-Ignited engine to capture the characteristics of a four-stroke engine.
This analysis will provide insights on spray parcel ejection, turbulence, combustion, fuel utilization, and engine properties variation at various stages of the power generation process. It will also provide information on the compression ratio, combustion efficiency, power and torque produced by the simulated engine to provide a sense of complete engine simulation.
Parameter Guide
*Total Mass Injected was calculated as represented in the table below
1. Introduction
Converge Studio is the flagship model from Convergent Science which is widely used for CFD application. With its advanced physical models and high-performance computing, it provides an easy to interact GUI and focuses on applications in IC engines, gas turbines, exhaust treatment, and general flow.
In this project, Converge Studio was used to simulate and analyze a Port Fuel Injected Spark-Ignited engine under certain parameters, which are elaborated systematically in the parameters guide.
Also, in this project, we will see the application of spray modeling, combustion modeling, turbulence modeling and spark modeling to replicate accurate functioning of the engine.
The flow of information in this report will be as follows
- Motivation
- Understanding
- Geometry
- Case Setup
a. Preprocessing
b. Solver
c. Postprocessing - Computational Analysis
a. Compression ratio of the engine
b. Combustion efficiency of the engine
c. Power and torque of the engine
d. Significance of crank angle. - Inference
- Reference
1. Motivation
It takes years to sense the impact of a single high-intensity drop in the ocean of global phenomenon. The naturally occurring flow of global properties faces disturbance when any artificially occurring flow of incident is introduced for a period of time. One such incident can be tagged with pollution; while there can be numerous sources to ignite such an incident, one can never ignore the source with one of the highest magnitudes of correlation to this incident i.e. automotive.
While automotive has been a boon to us humankind and will always be when it comes to moving us from point A to B, it has simultaneously instigated a hazardous environment to live in. Cars produce carbon monoxide, hydrocarbons, nitrogen oxide, and particulate matters which contributes big time not only to the deterioration of our planet but also to the deterioration in the health of humans. Pollutants from cars cause cancers, asthma, birth defects, heart diseases, and eye irritation to name a few.
While we can go on and on about the issues that automobiles make, one has to notice where it originates from. Engine, the heart of a vehicle, powers the vehicle to move in the desired direction and helps exponentially to reduce the load of traveling distant locations. An engine burns fuel to operate its parts so as to produce energy and power necessary to make the vehicle move, but an engine is never able to utilize all the chemical properties of the fuel in the process of power generation. The portion of the fuel that either remains off from utilization or goes unburnt releases from the exhaust of an engine to the environment. These after-products when coalesced produces carbon monoxide, hydrocarbons, nitrogen oxide, and particulate matters to name a few, thereby affecting adversely to our planet and everyone on it.
So why do we still use such a machine which causes such adverse effects? We all wish the answer to this question was as simple as the question itself. Engine, though, causes harmful effects has continuously been proven to help our civilization to foster and grow. It has constantly provided us with a better lifestyle and similarly, hundreds of thousands of reasons will come up to back up the viability of such a machine.
So does that mean it is inevitable from continuing to depreciate our planet? Well NO.
We cannot remove the machine but we can definitely help in improving it; improving it in every aspect of its functioning. So while we are at improving its power we do aim at achieving it without affecting our environment. So no, you may have a really powerful engine which might take you to point B from point A in a jiffy but it won't be actualized in doing so if it harms our planet. Scientists spend their lifetime in research pertaining to engine and they have been doing so for a century. Hence, that might say a lot about how crazy we are on keeping this mean machine, right? Hehe, yes we are but at the same time, we can't completely drive off the fact that engines are actually on a decline. Yes, not a headline but something to bring our motivational moment to a sweet paragraphical end, or does it?
By now we have learned that the electric motors are the savior of our planet. It does everything that our mean machine can do and does so by not affecting us or the planet in the way our mean machine does. The introduction to the electric motor seems to explain it all. So, it is true that with the advent and rise of the pure electric vehicles, internal combustion engines have started noticing that its days are numbered. Analysts have started auctioning on its exit year; few say 2030, some go with 2040 and some play it safe by calling 2050. As we do see an exit for our mean machines, it might not be as straightforward as it seems. Hybrid vehicles have been showing some promising performance when compared to pure electric vehicles, which leads to strong speculation of IC engines staying for good. Engineers, scientists, and enthusiasts are investing millions of dollars, if not billions, for this cause by constantly improving the machine, constantly increasing its efficiency and power delivery while continuously bringing its emission down; showing our mean machine too can propel a vehicle as efficiently as electric motors, if not more. The government too comes off to help as a friend in disguise by periodically regularizing stringent emission norms making players develop engines performing even more efficiently than the last one. This is where engine simulation just takes away the floor. Making the engine more efficient than it is, is a daunting task and requires tones of dollars to be invested to help fund the process. Engine simulation helps aid these issues. Engine simulation provides realistic data which can be used to model an efficient engine without having to physically run an engine iteratively to achieve the same output, saving quite a lot of the aforementioned tones of dollars. Engine simulation also serves in saving an ample amount of time which can be utilized in interpreting the next big step in making an engine perform even more efficiently. So to conclude, to get a sense of what an engine simulation looks like and how it is executed to achieve data of interest, let's dive into one.
2. Understanding
With the required motivation, we now proceed into the most interesting step of simulating an engine, but before doing so its necessary to comprehend a few concepts which aid in parametrizing the simulation.
In this project, a CAD model has been implemented in order to replicate a real-life engine functioning which includes injection, compression, combustion, expansion, and exhaustion. The CAD model has been developed using experimental data obtained from research.
Upon importing the model, the first step was to analyze the model for any geometry errors. Upon successful correction of these geometry errors, segregation of the geometry was executed in order to understand the engine functioning part by part. This step was followed by defining region based on engine functioning such as intake, combustion, and exhaust. This process can be categorized under geometry cleaning as it involves correction and flagging of appropriate items as required, based upon which the further simulation is to be performed.
Once, the geometry cleanup was performed, its time to parametrize its functioning in order to achieve outputs of interest. Converge Studio offers several tools such as Spray Modeling, Combustion Modeling, Turbulence Modeling, and Source/Sink Modeling to name a few to aid replicate the functioning of an engine. These extremely critical and helpful tools were utilized using experimental data to draw our simulation.
Spray Modeling
Spray modeling toolbox was utilized to form the mechanisms of the injector and the nozzles in order to achieve parcel delivery from the intake to the combustion chamber. The spray is a very critical topic of discussion and can be formulated to replicate the exact flow characteristics of fuel, but due to limitations in the computational power and time, usually, sprays are approximated to reflect a fairly accurate model to save computational cost; unless spray, in specific, is of interest in which case computational cost is not considered to be of concern. Converge uses the Lagrangian solver to model discrete parcels and Eulerian solver to model the continuous fluid domain, as shown in Fig. (a).
Fig. (a)[6]
We mentioned parcel delivery earlier in our discussion. To define parcel, Converge Studio utilizes the model idea where the spray field if seen under an exploded view, can be seen to have drop packets which if exploded further will be seen to contain smaller sized droplets. These drop packets are known as parcels, which contains droplets. Each and every droplet in one parcel have the same geometric and thermodynamic properties. These parcels statistically represent the entire spray field. The physical propagation of these parcels is atomized, as follows, to comprehend the fuel flow, also represented in Fig. (b) for clearer understanding.
a. Primary breakup
b. Secondary breakup
c. Collision and coalescence
d. Drop drag
e. Turbulent Dispersion
f. Evaporation
Fig. (b)[6]
In our case, an injector with four nozzles was utilized for simulation. The models utilized to mimic a real-life injector functioning are Kelvin Helmholtz model and Rayleigh-Taylor model complemented by a discharge coefficient model, details of which has been produced in the subsequent sections. The fuel used for our case is `IC_8H_18` - Isooctane, as it has properties that are very close to the actual gasoline fuel used in the engine. A characteristic plot for `IC_8H_18` - Isooctane is shown in Fig. (c).
Fig. (c)
Combustion Modeling
In the case of Spark-Ignited engines, fuel and air are mixed at a molecular level prior to ignition and combustion occurs as the flame front propagates into this mixture. The process can be divided into three stages as
a. Ignition Lag: Represents the stage with a time lag between the instant of spark and the instant of pressure rise due to combustion.
b. Flame Propagation: Represents the region of flame propagation through the mixture, once it becomes self-sustaining as the rate of heat generation by burning is greater than the heat lost by the flame to the surrounding.
c. After-burning: Represents the period where the combustion continues even after attaining peak pressure which takes place when the mixture is rich.
A representation of the three combustion stages is shown in Fig. (d).
Fig. (d) [1]
Converge Studio provides several combustion models which can be utilized for engine specific solver for the chemical kinetics in order to capture these stages. In our case, as we are dealing with a gasoline port fuel injected engine, detailed chemical kinetics model (SAGE), which utilizes CVODES solver, was used to model the combustion process.
Turbulence Modeling
To capture turbulence as accurately as possible, the RANS solver with RNG `k-epsi` turbulence model was used to represent efficiently accurate turbulence characteristics. Due to the limitation of computational power and time, Large Eddy Simulation (LES) solver wasn't used and hence, Reynolds-Averaged Navier Stokes (RANS) solver was used, which aids in attaining accurate and meaningful turbulence characteristics.
To know more about RANS solver and why we use it, please visit my page: https://projects.skill-lync.com/projects/Reynolds-Averaged-Navier-Stokes-equation-07693
Usually, RNG `k-epsi` and Rapid Distortion RNG `k-epsi` turbulence models are utilized for IC engine application. RNG `k-epsi` turbulence model, though is highly numerically sensitive, covers and accounts for more scales of motion than the Rapid Distortion RNG `k-epsi` turbulence model and hence, is able to capture the engine functioning more accurately. Rapid Distortion RNG `k-epsi` turbulence model is used where rapid compression and expansion is involved, which isn't required in our case. Also, Rapid Distortion RNG `k-epsi` lacks validation with theory and test cases. Hence, RNG `k-epsi` turbulence model was preferred over RNG `k-epsi` turbulence model.
Source/Sink Modeling
This option from Converge Studio was used to model the spark. Appropriate spark generation model is critical to capture the correct ignition and to avoid exhaustion of computational cost. The duration of spark, which is dependent on the energy of the spark needs to be model appropriately to gain appropriate simulation.
Wall Heat Transfer Model
Currently, most of the research is focused on the combustion process as this is the most complex region of the engine process. However, one critical topic in this is the inaccurate conversion of thermal energy to mechanical energy due to the limitation in estimating thermal loss through the combustion chamber wall. Reliable estimation of wall heat transfer is essential for the overall accuracy of engine simulation with respect to combustion efficiency, engine exhaust and component thermal stress [2]. Therefore, capturing this variation of temperature within the chamber is very critical to achieve accuracy. Calculating the wall heat transfer is computationally exhaustive and hence a wall heat transfer model is utilized in such simulations. A wall heat transfer model simplifies the user's input and reduces solving time.
As we approach the simulation process, it is necessary to define a mesh size for the engine so as to accurately capture information about its functioning. Mesh size needs to be decided based on computational capacity. In our case, a grid size of 0.004 m was used due to a limitation in computational power. However, adaptive mesh refinement and fixed embedding have been implemented in the regions of interest to capture accuracy.
Now, we have gathered a good amount of information required to parametrize the simulation and execute it for analysis. Hence, let's dive in.
3. Geometry
The above-attached geometry is a port fuel injected spark ignited engine model that was used for the simulation and analysis.
This initial geometry, when imported to converge, was diagnosed to have several errors which were required to be analyzed and removed for application of the engine variables and proceeding with the simulation. The errors and its corrected forms are displayed as follows
Open Edge Errors
Intersection Error
Normal Orientation Error
Upon correction of the errors, the engine was flagged with necessary boundaries and was categorized under interest-specific regions, as shown in the figures below, for easier interpretation and understanding of the engine construction as well as the data produced from the simulation.
4. Case Setup
a. Preprocessing
In this section, we basically set up the necessary operating parameters for our geometry. Upon transfer of the geometry, boundary conditions were defined on the respective faces of the geometry namely Piston, Liner, Cylinder Head, Intake Port 1 and 2, Intake Valve Top, Intake Valve Angle, Intake Valve Bottom, Exhaust Valve Top, Exhaust Valve Angle, Exhaust Valve Bottom, Inflow Outflow, Spark Plug and Spark Plug Terminal.
With the completion of the above process, the case was set up with the necessary parameters as mentioned in the Parameter Guide.
b. Solver
Once the setup was done, the files were exported to an allocated folder for processing. Processing was carried out using Cygwin, a DOS environment which extracts the information from the exported files to process the simulation and create an output folder, which was then used to create log files to perform graphical visualization of the simulation for analysis.
c. Postprocessing
The output files from Cygwin was then used to process a log file consisting of all the information of the simulation which will be required to produce a visual for the analysis. Converge Post-Processing 3D window helps to perform this task. However, both compatible Cygwin and Windows Command Prompt can also aid in performing the same task in the DOS environment. The advantage of using Converge Post-Processing 3D is that it provides a more interactive and easier environment to perform the task of unfurling and combining the information. Once the post-processing was performed, a group of ParaView compatible files was generated which was used to produce the graphical visualization. ParaView is a graphical user interface used to produce graphical outputs to the simulation performed on Cygwin using the data from Converge Case Setup.
First, a no-hydro simulation was executed to verify if the engine functioning including the valve propagation and timing were as per data. A no-hydro simulation is always advisable when dealt with complex geometry to avoid several errors that might arise upon simulation thereby preventing exhaustion of computational cost. The animation of the no-hydro simulation is attached below.
In the no-hydro simulation, the geometry shows the successful propagation of respective engine sections. Also, it can be noticed that Converge Studio has successfully implemented a restrictive triangular mesh around the valves to block the flow of fuel out of the respective region of interest.
Upon verifying for correct engine functioning, a full-hydro simulation was executed. The animation of full hydro simulation is attached below.
Upon simulating the PFI SI engine with full hydro under the above-mentioned parameters from the table, the mesh, temperature and pressure profiles were studied. The outputs from Cygwin were also used to understand the pressure, temperature, spray, and emission profiles using Converge Line Plotting window.
Mesh Profile
Initial Timestep
Timestep = 180 deg.
Timestep = 250 deg.
Final Timestep
It can be seen that Converge Studio has successfully manipulated the mesh size in the regions of interest, based on the parameters provided.
Liquid Spray Drop Profile
As the simulation was parametrized to introduce parcels during the intake phase, a surge in the parcel count can be seen at the intake cycle during the combustion process.
Property Profile
Emission Profile
The above plot shows the variation of four major types of emissions at various crank angles. These emissive constituents are the remainings of the fuel that went unburnt or incompletely burnt. These constituents are subjected to various after-treatment systems before been released in the atmosphere through the exhaust.
It can be clearly seen that after the end of the combustion process, the composition of Carbon Monoxide is higher than that of other constituents and this is generally caused due to incomplete combustion. Theoretically, in ideal cases, Carbon Dioxide is the after product of combustion but in real life when enough Oxygen is not present for the combustion process Carbon Monoxide forms. It can also be seen that the engine produces other constituents such as HydroCarbons, Soots and Nitrogen Oxides, constant research and efforts are made in the engine development process to bring these emissions down so as to meet stringent emission norms.
Timing Map
An overview of the complete engine process can be visualized in Fig. (e)
Fig. (e)
5. Computational Analysis
Upon generation of data based on the parameterized simulation, couple of analysis were carried out to understand the functioning and properties of the engine.
Engine performance (as produced by the simulation)
Abbreviation used in the table
CA: Crank Angle
IMEP: Indicated Mean Effective Pressure
CA10: Crank angle at which 10% of the fuel is consumed.
CA50: Crank angle at which 50% of the fuel is consumed.
CA90: Crank angle at which 90% of the fuel is consumed.
PFP: Peak Firing Pressure
PFP CA: Crank angle at peak firing pressure
HRR: Heat Release Rate
HRR peak CA: Crank Angle at peak heat release rate.
a. Compression ratio of the engine
The compression ratio of an engine is the ratio of the maximum combustion chamber volume to its minimum combustion chamber volume. In other words, the ratio of the volume of the combustion chamber when the piston is at its lowest position to the volume of the combustion chamber when the piston is at its topmost position. Compression ratios can be manipulated accordingly to improve fuel efficiency and economy.
From the simulation, the maximum and minimum swept can be obtained by plotting the volume vs crank angle, as shown in Fig. (e).
Fig. (f)
From the plot,
Maximum Swept Volume = 0.0005742195 `m^3` ; at Crank Angle = -176.23861 deg.
Minimum Swept Volume = 5.7029438E-05 `m^3` ; at Crank Angle = -360.22063 deg.
or,
or,
b. Combustion efficiency of the engine
It is the measurement of how well the supplied fuel is utilized in the combustion process. In other words, it is the ratio of heat released to heat input by the fuel.
From the simulation, the integrated heat released can be obtained by plotting the integrated hrr vs crank angle, as shown in Fig. (f).
Fig. (g)
From the plot,
Integrated heat released = 1241.1646 J
Amount of fuel supplied = `3\times10^{-5}\ (kg)/(cyc.)`
Calorific value of fuel supplied = 45 `(MJ)/k`
or,
or,
or,
c. Power and torque of the engine
While the power of an engine represents the maximum power an engine can put out [5], the torque of an engine represents the amount of turning power the engine can produce which helps in the rotation of the wheels in a vehicle.
We know, `Power = \frac{Wo rk}{Time}`
RPM = 3000
Work = 468.646 Nm
From the engine performance table, Duration = 240.199 deg.
Therefore, time per duration = `= \frac{60}{360\ times\ 3000} times 240.199 = 0.013344 s/(deg.)`
From the calculated time, `Power = \frac{468.646}{0.013344\ }=\ 35.12\ KW`
For Torque,
We know, `To rque = \frac{60\ times\ Power}{2\pi\ times\ RPM\ }`
or, `To rque\ =\ \frac{60\ times\ 35.12\ times\ 10^3}{2\pi\ times\ 3000}`
or, `To rque\ =\ 111.85\ Nm`
d. Crank angle at respective portion of fuel consumed.
From the engine performance table above, the following information can be gathered.
Crank angle at which 10% of the fuel is consumed: 6.83717 deg.
Crank angle at which 50% of the fuel is consumed: 18.4623 deg.
Crank angle at which 90% of the fuel is consumed: 31.701 deg.
6. Inference
Even though the CAD model was developed using experimental data, no output data is available for simulation validation. However, from the output and animation obtained from the simulation, it can be seen that the simulation is in accordance with a standard SI engine. The simulation has enabled us to visualize the intensive engine operating process and understand the various perspective of individual importance of each region as well as properties. Such a simulation is very crucial in establishing experimental validation and improvement aiming towards meeting stringent norms introduced by the government. With such simulation implementation, we can see how we can keep improving an engine with low computational cost and faster results without affecting the accuracy in the engine performance.
7. Reference
[1] https://www.ques10.com/p/6592/explain-in-detail-the-various-stages-of-combusti-1/
[2] Sanjin Saric and Branislav Basara; A Hybrid Wall Heat Transfer Model for IC Engine Simulations (April 2015).
[3] https://en.wikipedia.org/wiki/Engine_efficiency
[4] Engine Design and Development; https://www.iav.com/sites/default/files/handouts/2014/en-us/engine-design-and-development.pdf
[5] https://en.wikipedia.org/wiki/Engine_power
[6] Converge CFD Software; https://convergecfd.com/