An Investigation into the Combustion Characteristics of a SI Engine to Verify Simulation Models
Jack Lu
Letter of Transmittal
Abstract
Table of Contents
Acknowledgements
1. Introduction
2. Background
Today’s automotive industry heavily relies upon the use of engine simulation software to develop and design both race and conventional engines. Sophisticated one dimensional engine Computational Fluid Dynamics (CFD) packages such as RICARDO WAVE® cost millions of dollars but are capable of producing results within an error of 1-3% of dynamometer (DYNO) results1. Like any other CFD or Finite Element Analysis (FEA) software, the results produced are only as good as the data input.
UWAM has been utilising WAVE® to design their powertrain package since 2004 and over the past three years have defined their model to within 10% of their DYNO torque curve. The significance of acquiring such an accurate model for UWAM is so that the effects of powertrain hardware can be simulated during the design stage. Without an accurate engine simulation, multiple prototypes have to be manufactured and tested. In some cases large modifications to the engine itself can deem the resultant hardware ineffective and hence ultimately cost the team a large amount of time and finances.
Recent work produced by Lu, 2007, and advances in engine research within UWAM have enabled one to retrieve pressure information from within a cylinder of the Honda CBR600R. When this data is logged alongside with crank angle it is capable of producing a Pressure-Volume (P-V) plot that is undoubtedly invaluable to the understanding of the thermodynamic system of the engine.
3. Literature Review
Past UWAM theses and engineering projects extensively cover intake design and gas exchange for a Formula SAE car particularly naming Kitsios (2002), Inkster (2004), Kawka (2005) and Rogozinski (2006). Kitsios centred his work upon the theoretical fundamentals of gas dynamics and produced a basis for engine simulations upon which Rogozinski built his CFD simulations and experimental analysis upon. The core of Inkster and Kawka’s work revolved around the design of a variable runner intake plenum and proved that educated decisions can be made from full engine simulation models. Paget (2003) evaluated and utilised the Stanford Engine Simulation Software to design for and optimise engine valve train. His work included the use of an in-cylinder pressure transducer but high levels of noise with secondary cyclic signals superimposed upon the main signal were evident. These were concluded to be caused by the existence resonance occurring within the 30mm long bore between the piezoelectric sensor and the cylinder.
‘Design and Simulation of Four Stroke Engines’ by Prof. Gordon P. Blair covers engine simulation techniques in high detail with extensive references to the fundamentals of thermodynamics. Experimental research and case studies are provided to verify his models along with a great deal of advice on increasing engine efficiency and performance.
‘Internal Combustion Engine Fundamentals’ by John B. Heywood is an extensive review of the vast and complex mass of technical material that now exists on spark-ignition and compressionignition engines. Heywood comprehensively covers all aspects of gas dynamics related to the internal combustion engine by applying the laws of chemistry and thermodynamics. A great deal of Heywood’s work is backed up by experimental results and illustrations.
‘Measuring Absolute-Cylinder Pressure and Pressure Drop Across Intake Valves of Firing Engines’ by Paulius V. Puzinauskas, Joseph C. Eves and Nohr F. Tillman is a technical paper describing a technique which can accurately measure firing-cylinder full-load absolute pressure during intake events, thereby providing useful cylinder-pressure data for valve-timing optimisation.
‘Spark Ignition Engines – Combustion Characteristics, Thermodynamics, and the Cylinder-Pressure Card’ by Frederic A. Matekunas is a research paper covering the thermodynamics theory behind combustion and discusses about the factors that are important to the timing of the burn for maximum brake torque operation.
4. Combustion Process within the Four Stroke Cycle
An internal combustion engine gains its energy from the chemical energy released during the combustion of the fuel/air mixture and therefore the combustion process dictates engine power, efficiency and emissions. The combustion process of a four stroke spark ignition (SI) engine can be divided into four distinct phases: spark ignition, early flame development, flame propagation and flame termination. The four phases lie between the compression and power strokes of the four stroke engine cycle seen in Figure 1. During the intake stroke the piston falls from top dead centre (TDC) increasing the cylinder volume while the intake valve is open. A fresh charge of fuel/air mix is inducted through the intake valve and into the cylinder mixing with the residual gas that remains in the cylinder from the previous cycle. During the compression stroke all valves are closed and the cylinder volume decreases as the piston moves up from bottom dead centre (BDC) compressing the gas mix. The combustion process is initiated by the spark plug towards the end of the compression stroke under normal operating conditions and continues through to the early portion of the power stroke. At this point a turbulent flame develops and propagates through the fuel/air/residual gas mix away from the spark plug and towards the chamber walls before extinguishing. Upon the start of the power stroke the cylinder pressure increases significantly and work is transferred to the piston pushing it down towards BDC ultimately increasing cylinder volume. The exhaust valve opens before BDC and the exhaust stroke expels the exhaust gases from the rising piston leaving some residual gases behind.
Figure 1 The 4 stroke engine cycle
4.1 Spark Ignition
The ignition within an SI engine is provided by the discharge of the spark plug that is generally controlled by an electronic control unit (ECU). The spark ignition initiates the combustion process and therefore controls the burn.
4.2 Early Flame Development
The early fame development (EFD) stage comprises of the flame development process from the spark discharge which initiates the combustion process to a point where a small but measurable fraction of the charge has burned or fuel energy released. In industry it is common to indicate the end of the EFD stage when 10% of the charge mass has been burnt. Other figures such as 1% and 5% have been used also.
4.3 Flame Propagation
The flame propagation stage comprises of the rapid burning of the charge. During this stage each element of fuel/air burns and its density decreases by a factor of four. The expansion of the combustion product gas compresses the mixture ahead of the flame and displaces it towards the chamber walls. At the same time the already burnt gas behind the propagating flame is compressed and displaced towards the spark plug. Elements of the unburnt gas are of different temperatures and pressures just prior to combustion and are at different states after combustion and their condition is determined by the conservation of mass and energy.
4.4 Flame Termination
The flame termination stage comprises of the propagating flame reaching the chamber walls and extinguishing. At this point the combustion process has ended and a large portion if not all of the fuel energy has been released to produce work onto the piston. The amount of fuel energy released is dependent upon the efficiency of the expansion in burn.
5. Variables Effecting Combustion
5.1 Combustion Phasing
Combustion events can be phased by advancing or retarding spark before top dead centre (BTDC). The phasing of the combustion event influences the magnitude and location of peak cylinder pressure by changing the rate of pressure rise within the chamber. Figure 2 illustrates the effects of combustion phasing by spark advance upon cylinder pressure.
Figure 2 Combustion phasing by advancing spark timing
By phasing the combustion so that the 50% mass burned point is closer to TDC allows complete combustion at TDC and therefore increases the compression stroke work transfer from piston to cylinder gases resulting in higher cylinder peak pressure. Ultimately this leads to increased work transfer from the cylinder gas to piston upon the power stroke increasing the brake torque output. Matekunas, 1984, introduces the idea of “phase loss” defined as the loss in efficiency as the 50% mass burned point is moved away from TDC. The optimum phasing that provides maximum brake torque (MBT) is known as the MBT point and any timing advanced or retarded from this point increases the “phasing loss” and produces lower torque. MBT phasing often produces a peak pressure location within the range of 14-17deg ATDC. Matekunas (1984)
5.2 Cylinder Turbulence
The combustion process in a SI engine occurs in a turbulent flow field. This flow field is produced by the high shear flows generated by the intake jet and flow pattern. In turbulent flows, the rate of transfer and mixing are several times greater than the rates due to molecular diffusion [Heywood (1988)]. One method of adding turbulence within the combustion chamber is known as squish action and this is caused by the geometry of the combustion chamber as the piston rises and compresses the gas. Squish characteristics in SI engines are relatively moderate compared to that of a compression-ignition engine. Another method of promoting turbulence is through swirl and tumble caused by the intake geometry.
5.3 Swirl and Tumble
The terms ‘swirling’ and ‘tumbling' are used to describe the rotating of flow within the cylinder. Swirl is defined as the controlled rotary motion of the charge about the cylinders axis whereas tumble (Figure 3) is in cylinder flow at right angles to the cylinder axis. They are created by providing an initial angular momentum to the charge as it enters the cylinder through the intake ports. Swirl and tumble can assist in speeding up the combustion process within SI engines and hence achieve higher thermal efficiency.
Figure 3 Tumbling of the intake charge within the cylinder
Measuring Swirl and Tumble
Swirl ratio
5.4 Compression Ratio
The compression ratio (CR) is defined as the ratio of maximum volume (when the piston is at BDC) to minimum volume (when the piston is at TDC). At BDC the volume comprises of the swept volume Vs and the clearance volume Vc whereas at TDC the minimum volume at which combustion occurs consists of only the clearance volume Vc.
CR=maximum volumeminimum volume=VbdcVtdc=Vs+VcVs
In the APPENDIX Blair (1999) proves that the highest thermal efficiency is achieved at the highest compression ratio but if the compression ratio is too high, engine operation will exhibit abnormal combustion which is an undesirable outcome.
6. Abnormal Combustion
Normal combustion is initiated by the discharge of the spark plug and develops a flame that propagates to the chamber walls before extinguishing but there can be several factors to cause abnormal co mbustion. These factors are fuel composition, engine design and operating parameters and combustion chamber deposits (Heywood, 1988). The two most common forms of abnormal combustion are identified as knock and surface ignition. Both of these reduce the combustion efficiency and through persistence will destroy engine components by exceeding the engines pressure design limits. Figure 4 illustrates the difference normal and abnormal combustion as seen from a pressure trace.
Figure 4 Pressure trace involving abnormal combustion
6.1 Knock
Knock is described as the sharp metallic noise caused by the auto-ignition of the fuel/air/residual gas mix ahead of the propagating flame. During combustion the propagating flame compresses and displaces the end gas ahead of the flame towards the chamber wall. This causes its pressure, temperature and density to increase undergoing the chemical reactions prior to normal combustion. When pressures and temperatures become excessive the end gas burns very rapidly releasing a large amount of its energy at a rate five to twenty five times normal combustion causing high frequency pressure oscillations within the chamber that exceed engine design limits. These detonations are initiated by high pressures and temperatures and therefore can be avoided by reducing the compression ratio, using a higher rating octane fuel, appropriate calibration of the engines ignition timing and careful design of the engines cooling system.
6.2 Surface Ignition
Surface ignition is the uncontrolled ignition of the fuel/air/residual gas mix from overheated valves, walls, spark plug or glowing deposits. There are two types of identifiable surface ignition: pre-ignition and post-ignition. Pre-ignition can be identified from a pressure trace as the combustion event is initiated before the targeted spark ignition time and causes the most severe effects as the spark no longer controls the combustion process. Post-ignition occurs after the spark ignition but can be difficult to distinguish from knock as they both portray the same characteristics under a pressure trace.
6.3 Cyclic Variations
It is evident from observation of cylinder pressure versus crank angle measurements over consecutive cycles within a sample that cyclic variation exists. For a motored pressure trace cyclic variations are negligible and pressure measurements tend to follow closely to the polytropic relationshippVn=constant. Therefore the pressure development is distinctively related to the combustion process which is dependent upon different variables. These cyclic variations are caused by variations in charge motion and mixture motion at the time of the spark, the amount of fuel/air within the cylinder and the fuel/air ratio, and the mixing of the fresh mixture with the residual gases remaining in the cylinder. Along with cylinder cyclic variations there also exists cylinder to cylinder variance in multicylinder engines which are caused by the same reasons.
It is important note that due to cyclic variations, the optimum combustion phasing will then be different for each variation of combustion and that the MBT spark advance experimentally found from engine tuning methods described by Bleechmore (2006) is set for the average cycle. Any cycles faster than the average cycle effectively advances spark timing away from the MBT point and any cycles slow than the average cycle effectively retards spark timing away from the MBT point.