Effect of elevated piston temperature on combustion chamber deposit growth full repor
Effect of elevated piston temperature on combustion chamber deposit growth ppt.PPT (Size: 485.5 KB / Downloads: 220)
Effect of elevated piston temperature on combustion chamber deposit growth report.doc (Size: 734.5 KB / Downloads: 170)
Combustion chamber deposits are found in all internal combustion engines. When fuel and air burn inside an internal combustion engine, deposits form on the walls inside the cylinder. The factors influencing deposit formation are changes in fuel composition, coolant temperatures, engine speed and load, and spark timing. One way to discourage deposit formation is to prevent condensation. Raising the wall temperature prevent condensation. With increased wall temperature, hydrocarbon condensation tendencies decreases and thus deposit formation is reduced.
In these attempts is made to investigate the effects of wall temperature on deposit growth, without the use of ceramic coatings on the combustion chamber surfaces. Also methods for monitoring combustion chamber deposit growth as a function of metal wall temperature are described. In this, a composite piston design was developed to control and monitor surface temperature. The design incorporated ceramic wafers of varying thickness encapsulated between a metal cap and piston top. The cap was fitted with surface-mounted thermocouples to measure wall temperatures. An accelerated test cycle was used to accumulate deposits. There is a description of methodology employed in raising the wall temperature and monitoring deposit growth. After stabilization of deposit growth a physical and chemical analysis of deposits from different locations were also conducted.
Combustion chamber deposits are recognized as a major contributor to the deterioration of SI engine performance . Their build-up leads to reduced air mass flow rates, increased charge emissions, and increased tendency for knock. The primary mechanism of formation of fuel deposits is condensation of the high boiling point components, such as aromatics, and their carbonization while in the liquid phase. The amount of air mixed with the fuel has been found to be critical to the deposit growth in engines. Therefore, oxidative fuel pyrolysis better describes the mechanism of deposit formation. With increased wall temperature, hydrocarbon condensation tendencies decrease, and therefore deposit formation is reduced. Deposits on hot surfaces, such as the exhaust valve of the spark plug, are primarily composed of oil-based inorganic compounds. Deposit growth in the end gas region, the region with the coolest surfaces, is typically higher than in any other portion of the combustion chamber, end chamber deposits also have lower oxygen and higher carbon content than elsewhere in the combustion chamber.
The purpose of this experiment is to investigate the effects of wall temperature on deposit growth, without the use of ceramic coatings on the combustion chamber surfaces. This study monitors in -situ combustion chamber deposit growth, as a function of metal wall temperature, and attempts to determine the critical wall temperature for no growth a composite piston design was developed to control and monitor surface temperature. The design incorporated ceramic wafers of varying thickness encapsulated between a metal cap and piston top. The cap was fitted with surface-mounted thermocouples to measure wall temperatures. An accelerated test cycle was used to accumulate deposits. This also contains the methodology employed in raising the wall temperature and monitoring deposit growth.
COMBUSTION CHAMBER DEPOSITS
Combustion chamber deposits are found in essentially all internal combustion engines. Their influence can be seen in many aspects of engine performance. Deposits form on all engine surfaces that are in contact with fuel or fuel air mixture at any point in the cycle. The mechanism by which these deposits form from fuel are not well understood .the effect of deposit include increased engine NO emissions, octane requirement increase and changes in flame speed and thermal efficiency.
The factors influencing deposit formation are changes in fuel composition, coolant temperatures, engine speed and load, and spark timing. These factors affect the rate of deposit formation and removal and the structure of deposits. To create a deposit control strategy needed is the magnitude and direction of these various effects over a wide range of fuel types and engine conditions.
The chemical and thermo physical properties deposits depend on fuel and oil composition, wall temperature, engine operating conditions, and engine design. The carbon to hydrogen (C/H) ration of the deposits indicates the relative proportion of fuel and oil contribution to the deposits. A low carbon to hydrogen ratio indicates deposit constituents are mostly oil based, and originate from oil cracking on the wall. A high carbon to hydrogen ratio indicates the deposits are primarily fuel derived. In general, deposits collected from engines that operate under normal conditions are primarily fuel derived.
OCTANE REQUIREMENT INCREASE
The actual octane requirement of a vehicle is called octane number requirement. A new engine require a fuel of octane number 6-9 lower than the same engine running for a long time. This octane requirement increase is due to the formation of a mixture of organic and inorganic deposits resulting from fuel and the lubricant. When using unleaded fuel two-thirds of the octane requirement is caused by deposit accumulation in the end gas region.
EFFECT OF COMBUSTION CHAMBER DEPOSITS
The effect of combustion chamber deposits include octane requirement increases, decreased volumetric efficiency, increased thermal efficiency due to insulation of the cylinder, combustion chamber deposit interference(CCDI) and combustion chamber flaking.
As deposits build up on the combustion chamber surfaces, the minimum fuel octane number needed to avoid engine knock increases. In modern engines this increase is approximately 4 to 5 octane numbers on average as the deposits build up to their equilibrium thickness. Combustion chamber deposit interference is the result of physical contact between deposits on the piston top and cylinder head. Combustion chamber deposit flaking causes low compression pressures due to improper sealing of the valves.
EXPERIMENTAL SETUP AND PROCEDURE
THE ENGINE CONFIGURATION
The investigation was performed using a Waukesha, single cylinder, split head, variable compression ratio, Cooperative Fuel Research (CFR) engine. Engine motoring and loading was handled by a General Electric DC dynamometer with 20 hp motoring and 35 hp absorbing capacity. The CFRâ„¢s intake manifold was modified to incorporate a throttle body in order to regulate engine load. An active magnetic pickup, mounted in close proximity to the dynamometer shaft, was used to measure engine speed. The dynamometer was controlled with a Dyne Systems DYN-LOC IV controller
The fuel delivery system was converted from the standard CFR system to a modern electronic fuel injection setup which was comprised of an automotive fuel fitter and pump, along with a Bosch fuel injector and pressure regulator. The fuel injector was installed in the in take manifold where the CFR injector was normally fitted. The ignition system consisted of a single pole, ignition coil and a TFI module that was used to trigger coil discharge. A fuel/spark controller was developed to control timing of the fuel injector and initiate coil discharge so as to achieve desired fuel flow rate and spark timing.
Engine oil and water jacket cooling was provided by tube type heat exchangers. The jacket water consisted of a mixture of approximately 80 percent water and 20 percent ethylene glycol. Engine oil temperature and jacket water temperatures were measured with K-type thermocouples. During testing, oil temperature in the crankcase sump was maintained at 90Ã‚Â± 20 Ã‚ÂºC. The jacket water temperature stabilized to 100 Ã‚ÂºC in 15 minutes after starting the engine.
Air flow was measured with an ASME 5/8 inch metering orifice with flange mounted pressure taps. Fuel flow rate was measured with a calibrated burette Cylinder pressure was measured with a water cooled piezoelectric transducer mounted in the knock sensor hole of he head. The transducer was cleaned and covered with high temperature silicone RTV.
CONTROL AND MEASUREMENT OF PISTON SURFACE TEMPERATURES
Deposit growth was investigated on the top surface of the piston, which typically accounts for two-thirds of the deposit growth in the combustion chamber.A composite piston design was developed to control and monitor surface temperature. The design incorporated a ceramic wafer between an added iron cap and the standard iron piston top. The ceramic wafer was fabricated from a glass/ceramic material with low conductivity of 1.38 W/m. Ã‚ÂºC, and high compression strength of 344 Mpa, and excellent machinability. The wafer measured 76.2mm in diameter, and 1.0, 2.0, or 3.0 mm in thickness. By varying the thickness of the ceramic wafer, the surface temperature of the cap could be increased from its baseline value. Predictions ware generated through a one-dimensional heat transfer analysis of the composite piston design, based on an assumed transient heat flux on its combustion surface.
A series of caps was machined from ferromagnetic iron, with each cap having a matching pocket to a seat a given ceramic wafer. For all caps, the distance between the top surface and its pocket was maintained at 4 mm for structural support. The cap was secured to the standard piston by countersunk fasteners. The CFR cast iron piston was used because its material closely matched the piston cap material. Since the iron piston has four compression rings and only one oil ring, the connecting rod oil jet, which is normally used for cooling the pistonâ„¢s underside, was sealed off to avoid excess oil consumption in the combustion chamber.
Surface-mounted thermocouples were mounted flush with the cap surface to allow monitoring of its temperature during engine operation. Four constantan wires were affixed to the cap at 36.5 mm radially from the iron wire. The J-type thermocouple wire used was 24 gauge and has a peak temperature range of 700 Ã‚ÂºC. The wires were fitted in the cap through counter-bored and counter sunk holes. The wires were bonded to the cap by welding the wire end into the countersunk holes.
TRANSMISSION OF ELECTRICAL SIGNALS-
Due to the harsh environment of high accelerations and temperatures, the piston assembly was meticulously wired to insure electrical continuity between the capâ„¢s thermocouples and the connecting road. Augat electrical connectors which forming good electrical contact with 24-gage copper wire, were placed in the piston top to serve as an electrical socket for the thermocouple wires. Standard 26-gauge copper wire, which was secured to the connector with silver solder, was used to continue transmission of the electric signals. An RTD was placed next to a protruding Augat connector on the pistonâ„¢s underside to establish the reference junction temperature. The RTD and all wires were bonded to the piston with epoxy.
A stainless steel band was used to provide a path between the piston and the connecting rod. The band was mechanically secured to the pistonâ„¢s skirt and the connecting rod. Finely stranded copper wires were guided along the bandâ„¢s outside surface and held with high-strength RTV and segmented high-strength heat shrink tubing. After passing under two aluminum plates which were fastened to the connecting rodâ„¢s shank, the wires were eventually routed to the end cap of the connecting rod. In order to establish electrical continuity of the thermocouple signals from the end cap of the connecting rod to an external data acquisition system, a mechanical telemetry linkage system was designed. The system works on the concept of classical four bar linkage. The longest possible wire life was achieved by minimizing angular deflection and by maximizing the transmission angle between adjoining linkage members. This reduced bending of wires at linkage joints, thus promoting longer wire life.
A transient test cycle, appropriate for a single-cylinder engine, was designed for the test The test schedule was broken down into four and one half hour test cycles, followed by engine cool down to ambient temperature. Within each test cycle, a repeated fifteen minute test segment was conducted. One and one half minutes were spent at idle, thirteen minutes at low load, and one half minute at a high load. At the beginning of each test, the engine was warmed up at 800 rpm for twenty minutes.
The four and one half hour test cycle was repeated as many times as needed for establishing well defined growth patterns. For each four and one half hour test cycle, fast response, pressure and temperature data were recorded every half hour during the 1800 rpm segment. Additional measurements included airflow, oil and cooling water temperatures, exhaust temperature, and intake pressure. Fuel consumption measurements were documented periodically during testing.
At the end of testing with a given cap, the cylinder head was removed, and the deposit thickness on the head, valves, bore, ports and piston were recorded. Deposits were removed from the cylinder head, valves and ports. Then the valves were lapped and the engine was reassembled for further testing. All piston configurations were tested with a base premium unleaded fuel without reformer bottoms. The 3 mm cap test was repeated with the unleaded fuel doped with reformer bottoms.
IN-SITU MONITORING OF DEPOSIT GROWTH
A technique for in-situ monitoring of deposit growth is through measurement of local surface temperature using thermocouples affixed to the combustion surface. As deposits build up on the surface and form an insulation barrier, heat flow is reduced from the combustion gases to the wall. The restriction in heat flow reduces the wallâ„¢s surface temperature which is measured by the surface-mounted thermocouple. The rate of change of wall surface temperature is thus indicative of the rate of deposit growth. The surface temperature has been found to decrease almost linearly until deposit growth has stabilized, from then on, the surface temperature remains relatively constant.
Temperature gradients and stabilized growth periods have been found to be greatly influenced by the fuelâ„¢s hydrocarbon structure. Fuels with aromatic hydrocarbons, toluene and xylene, measured higher surface temperature gradients, during the first three hours of testing, than fuels containing paraffins, like iso-octane. The higher the boiling point of aromatic components, the faster the deposit growth, and the lower the measured wall temperature
Initially the baseline cap was run for a 42 hour deposit growth test. Recorded temperature histories exhibited two distinct temperature discontinuities at the 10th and 22nd hour. The first temperature discontinuity was due to the thermocouple wires and connectors being coated with a carbonaceous film, as a result of combustion gases penetrating into crevices of the contact surface between piston and cap. The gap was seated with a thin layer of high temperature RTV applied on the lower surface of the cap. The second temperature discontinuity was associated with vibration in the flat head fasteners. This was corrected by the use of flat head lock washers with the fasteners. After this, the first 12hours of the baseline test were repeated.
Regional differences in the measured rate of temperature decay are attributed to differences in deposit composition and thermal properties at the different locations. There is an accelerated deposit build-up in the regions surrounding thermocouples 3 and 4 (under the intake valve). The differences are driven by regional variations in wall surface temperature and changes in the C/H composition of deposits. The surface-averaged wall temperature of the baseline cap decreased linearly over time, at a rate of 0.38oC/hr.
EFFECT OF ELEVATED PISTON TEMPERATURE ON DEPOSIT GROWTH
TEST USING UNLEADED FUEL
Caps insulated with a1, 2 and 3 mm ceramic wafer were tested with the base unleaded fuel without reformer bottoms. An overall test period of 18hours was used to define deposit growth patterns. Local and averaged surface temperature histories of the insulated caps are compared with those of the baseline cap. Insulating the piston cap raised the initial temperature of the surface from an average of 215 Ã‚ÂºC to 317 Ã‚ÂºC for 3 mm cap .Elevating the initial temperature of the cap reduced the rate of decay with time, from â€œ0.38 Ã‚ÂºC/hr (baseline) to 0.02 Ã‚ÂºC/hr for the last 11hours of the test with the 3mm insulated cap. This shows an overall decrease in deposit growth with elevated piston temperature
The discontinuity in the surface temperature gradients of the 3mm cap is that light engine knock was experienced after 6hour of testing with the unleaded fuel without additives. Spark timing was then retarded by three crank angle degrees to eliminate knock, and the test was continued. A thermocouple histories in the 3mm insulated cap shows surface temperatures to decrease at a rapid rate, particularly in the end gas region of thermocouples 2 & 3, between the second and sixth hour. This decrease indicates accelerated early deposit growth, with the largest concentration in the end gas region. Deposit growth causes an increase in gas temperature, which leads to knock. The high thermal stresses crated by the knocking condition cause deposit flaking, or random removal of deposits. Once flaking occurred, deposit accumulation on the exposed surfaces was retarded. A deposit film remained on the capâ„¢s surface leads to periodic reoccurrence of knock and flaking, this shows the observed fluctuations in surface temperature after the sixth hour of testing.
There is a slightly positive slopes for the recorded temperature change at locations 2 and 3. This is attributed to , an inability to take data during the first one hour of testing, when a relatively large deposit accumulation occur at those locations, and difficulties during the first four-and-a- half hour test cycle. Data with the second test cycle establish negative (or close to zero) slopes. A little deposit growth is experienced with the 2 mm insulated cap, by an average rate of temperature decay of 0.06 Ã‚ÂºC/hr. With further increase in cap insulation, the mean temperature gradient during the last seven hours of testing was â€œ0.02 Ã‚ÂºC/hr, indicating negligible deposit growth with the 3mm insulated cap. An average wall temperature of 320 Ã‚ÂºC identifies a critical point above which the base unleaded fuel does not form deposits.
Results through in-situ monitoring of deposits using unleaded fuel yielded a critical surface temperature of 3200C.
TESTS USING FUELS WITH REFORMER BOTTOMS
Reformer bottoms are large hydrocarbon molecules which have poor oxidation characteristics, and thus enhance deposit growth rate. This is also affected by potential changes in the thermal properties of fuel deposits.
Testing with the reformer bottoms fuel was conducted for only eleven hours. The average wall temperatures for the reformer bottoms test were compiled from thermocouples 1 and 3.No knock occurred during testing of the 3 mm insulated cap with the reformer bottoms fuel. The linear decay in the average temperature of the piston surface (at a rate of 0.350 C/hr) shows a steady deposit growth. Reformer bottoms can promote deposit growth under operating conditions at which no deposits are formed with the base fuel.
The wall surface temperature for the reformer bottoms test was approximately 8 Ã‚ÂºC lower than that for the base fuel test, after one hour of testing. This difference can be accounted for by a 4 o C lower inducted air temperature for the reformer bottoms test. The other differences are a result of differences in thermal properties of deposits from the two fuels. The deposit layer from the reformer bottoms fuel experiences a lower temperature swing during the cycle, and a lower peak surface temperature, this also leads to a lower mean wall temperature and thus increased deposit growth. The deposit growth is effectively controlled by the temperature at the deposit surface. The temperature difference (5 0C) between the peak temperatures at the gas-exposed surface of deposits from unleaded fuels with and without reformer patterns, shows differences in deposit growth patterns.
TEST ANALYSIS OF DEPOSITS
A physical analysis of deposits is important in fully characterizing deposit growth patterns. At the end of each test schedule, overall observations were made concerning the color and texture of deposits in each of the four cap quadrants. In addition deposit thickness measurements were made with a Titan ZDM-1 microscope. This information was useful in correlating deposit growth and change in wall restricted within a square region defined by the thermocouple locations. The exact locations on the cap surface where deposit thickness measurements were taken.
Deposit thickness measurements are taken for various caps. At the baseline 12 hour test deposits in cap regions 1, 2, and 4 appeared as dark down in color with a sooty texture, while region 3 deposits had a lighter shade of brown and a crusty texture. After 18 hours of testing, the 1 mm insulated cap was covered with a deposit coating of a lighter shade of brown. In regions 1 & 4, deposits were observed to be denser and more sooty than those in regions 2 & 3, the regions 2 & 3 had a gritty texture. Compared to the 1 mm insulated cap which collected an average deposit thickness of 6.28 m, the 2mm cap had a reduced deposit growth (1.83 Ã‚Âµm on average), particularly in region 2 where no deposits had accumulated . The deposits on the remaining three quadrants of the cap had a light- brown metallic color with a gritty texture. The 3mm cap exhibited almost negligible deposit accumulation after 18 hours of testing. In regions 3 & 4, the cap surface seen under a rough and gritty texture whereas regions 1 & 2 were covered by a thin, smooth, brownish white deposit film.
The exhaust valve maintained a white color in all cases, at the end of each test with the unleaded fuel (without reformer bottoms).The intake valve color changed from a dark brown to a light gray as the piston surface temperature increased. At the end of the 3 mm insulated cap test, the exhaust valve had a clear evidence of deposit flaking. This is due to high surface temperature as well as engine knock.
The average deposit thickness was found to decrease at 3.38 Ã‚Âµm per millimeter of ceramic insulation. Low temperature regions, such as under the intake valve, were not always found to be covered under the thickest deposit coating. Thicker deposits accumulated on the cap for the longer test period. After 42 hours more deposits were found under the intake valve (region 3). Deposit flaking under the intake valve, where there is a tendency for large accumulation of deposits, resulted in removal of excessive deposit build-up with a long test period.
After 11 hours of testing with the reformer bottoms fuel, the 3mm cap accumulated a deposit layer with an average thickness of 0.47 Ã‚Âµm, 55% more than the layer accumulated after testing the unleaded fuel without additives for 18 hours. Aside from the poor oxidation properties of the reformer bottoms, the differences in these growth patterns are attributed to differences in thermal properties of the two types of deposits. Deposits from the reformer bottoms fuel were primarily centered in regions 3 & 4, as well as the edges of regions 1 & 2. They were identified with a dark brown color, and a rough, sooty texture. The intake valve from the reformer bottoms test accumulated a dark brown deposit layer. The exhaust valve appeared to maintain a light brown appearance. Deposit flaking was not found on this component.
Insulating the piston cap had a noticeable impact on deposit characteristics. The deposits were found to progress from a dark brown color with a sooty texture ( baseline cap ) to light-brown pigment with a gritty texture (3 mm cap ) Deposit growth on the cylinder head, showed similar trends. Deposits on the 3 mm insulated cap from the reformer bottoms fuel test exhibited comparable characteristics to those formed on the baseline cap using unleaded fuel without reformer bottoms.
A carbon to hydrogen (C/H) ratio analysis indicate the relative proportion of fuel and oil in the deposits, as well as the mechanism of deposit formation. The analysis reveals that deposits in the end gas region have the highest fuel content, as a result of deposition of unburned hydrocarbon products on the end gas region surfaces. The region between the spark plug and the intake valve has lower fuel content than the end gas region, and higher oil content. Since the former region runs approximately 16 Ã‚ÂºC hotter than the latter region, this behavior can be attributed to fuel vaporization on the higher temperature surfaces. The deposits collected near the capâ„¢s edge in the end gas region present the highest oil content. Excess oil on the cylinder liner accumulates on the latter region of cap during the pistonâ„¢s travel.
Mole fraction of carbon and hydrogen in the deposits collected under the exhaust valve portion of the end gas region (region 2) for each cap configuration tested with unleaded fuel without additives, shows an increase in the concentration of carbon within the deposits. The C/H ratio decreases with increased surface temperature. As wall surface temperature elevate, more oil contributes to deposits. A decreasing C/H ratio with increasing wall temperature indicates that deposit growth primarily originates from oil cracking on the wall. Deposits collected on high temperature, combustion chamber walls were composed mostly of inorganic compounds, such as those preset in engine lubrication. Deposits collected on cooler surfaces have been composed primarily of carbon.
CORRELATION OF TEMPERATURE DECAY AND DEPOSIT GROWTH
An analysis was performed that, measured changes in temperature over a given test period were normalized with respect to the measured deposit thickness at the corresponding location. The all four thermocouple locations were considered for the base unleaded fuel tests with the baseline metal cap (for 12 hours) and the 1mm insulated cap (for 18 hours). For all except one case, there appears to be a strong correlation between measured temperature decay and associated deposit thickness.
The effective conductivity of the deposits collected at a given location is nearly constant during the accumulation period. This holds true as long as there is no drastic change in local composition of deposits while increasing piston operating temperature. For the case of the 1mm cap, the region between the spark plug and the intake valve (region 4) progressively accumulated deposits of lower fuel content and higher oil content, with increasing piston temperatures.
When fuel and air burn inside an internal combustion engine, deposits form on the walls inside the cylinder. Deposits form on all engine surfaces that are in contact with fuel or fuel-air mixture at any point in the cycle. These deposits have significant effects. The deposit layer on the cylinder wall prevents the coolant that runs through the engine from bringing down the temperature inside the cylinder. Due to the increased temperatures, parts of the fuel-air mixture may ignite before the flame front reaches them, causing the engine to knock. Another effect of the raised temperatures is increased production of pollutants such as nitrogen oxides inside the engine cylinder. One way to discourage deposit formation is to prevent condensation. Raising the wall temperature prevent condensation. With increased wall temperature, hydrocarbon condensation tendencies decreases and thus deposit formation is reduced. Deposits on hot surfaces such as exhaust valves or ceramic of the spark plug are composed of oil based inorganic compounds. Deposit growth in the end gas region (region with coolest surfaces) is typically higher than in any other portion of combustion chamber.
Disadvantage of increasing wall temperature is that, if defects the overall goal of keeping cylinder temperatures down.
Combustion chamber deposits are found in essentially all internal combustion engines. When fuel and air burn inside an internal combustion engine, deposits form on the walls inside the cylinder. Factors that influencing deposit formation are changes in fuel composition, coolant temperatures, engine speed and load, and spark timing. These affect the rate of deposit formation and removal and the structure of the deposits. Deposits form on all engine surfaces that are in contact with fuel or fuel-air mixture at any point in the cycle. These deposits have significant effects. The deposit layer on the cylinder wall prevents the coolant that runs through the engine from bringing down the temperature inside the cylinder. Due to the increased temperatures, parts of the fuel-air mixture may ignite before the flame front reaches them, causing the engine to knock. Another effect of the raised temperatures is increased production of pollutants such as nitrogen oxides inside the engine cylinder.
One way to avoid deposit formation is to prevent condensation. Raising the wall temperature prevent condensation. With increased wall temperature, hydrocarbon condensation tendencies decreases and thus deposit formation is reduced. The unleaded fuel would not condense into a carbonaceous film once a critical wall temperature is reached. Reformer bottoms yielded a 55 percentage increase in deposit thickness compared to base unleaded fuel. This difference was due to poor oxidation properties of reformer bottoms. Elevating wall temperature increases the carbon to hydrogen ratio in the composition of deposits. Deposits in the end gas region are mostly fuel derived.
Disadvantage of increasing wall temperature is that, if defects the overall goal of keeping cylinder temperatures down.
Christopher Forssen oâ„¢Brien , Combustion chamber deposit research.
Energy laboratory research And related activities at the MIT , Inside engine cylinders; cleaner walls for lower emissions, higher efficiency
Chevron U S A , Gasoline vehicles deposit control.
The engineering society for advancing mobility land sea air and space, Effect of elevated piston temperature on combustion chamber deposit growth SAE Technical paper series ,940948,1994