Optical Engine Operation to Attain Piston Temperatures Representative of Metal Engine Conditions.
Improvement in thermal efficiency and a reduction in criteria emissions continue to be major challenges to the IC engine community. Optically accessible engines have proven to be a powerful tool for investigating in-cylinder topics such as air flow motion, fuel spray, ignition, fame propagation, and combustion product speciation . In these engines, a transparent window, typically made of quartz or sapphire, is installed in the cylinder head, cylinder wall, or employed as a piston insert. Aluminum, commonly used as the production piston material , has a higher thermal conductivity than the sapphire used in optical engines. Kashdan and Thirouard  measured the piston surface temperatures of both engines: optical and an equivalent all-metal engine. From these measurements, a skip-firing strategy was developed to compensate for differences in heat transfer properties in their optical engine.
An essential requirement to understand the combustion events when employing optical engines is to create thermal boundary conditions that are representative of the metal engine. Performing optical engine experiments before reaching metal engine representative temperatures can result in injecting the fuel into a low piston temperature environment. A lower piston window temperature hinders fuel vaporization and aggravates wall impingement and fuel film build-up, possibly resulting in higher soot formation in the optical engine. This notion was the motive behind developing the current temperature measurement technique. In addition, the measured optical piston temperatures can be useful for boundary conditions employed in fuel spray modeling. Previous temperature measurement techniques to determine surface temperatures of different engine components (piston rings, top land, crown and liner) can be classified into two types: Sensor measurements and radiation-or photon-based measurements.
The most widely used temperature sensor for engine applications is the thermocouple. Steeper and Stevens  instrumented surface thermocouples on a quartz piston and determined a skip-firing sequence that allowed a target temperature range of 145-160[degrees]C as determined separately in an all-metal engine. Thermocouples used for piston temperature measurements in the past had a slow response time . This problem was overcome by the fast response J-type and K-type NiCr-Ni thermocouples that offered the ability to measure surface temperatures with 0.3 [micro]s response time [6, 7]. For an engine speed of 2000 rpm, this response time corresponds to 0.0036 CAD. Another daunting task in using thermocouples is to lead the signals out of the piston to the data acquisition system. Addressing this problem, several designs have been developed based on two major mechanisms: mechanical linkage wiring systems and wireless mechanisms. Linkage wiring systems have been used in a number of piston heat transfer studies [8, 9, 10]. Such systems, however, are noted to have short lifetimes due to cyclic bending of the wires leading to wire fatigue and breakage . Alternatively, thermocouple signals can be transferred through "wireless" mechanisms such as electromagnetic induction [12, 13] or Bluetooth technology [14, 15]. However, higher accelerations in a piston-cylinder assembly make the wireless systems complex to install and prone to short lifetimes.
In addition to durability issues with thermocouples and wires, combustion deposits that can accumulate on the piston surface tend to bury the thermocouple junctions. As a result, the measured instantaneous temperature profiles change as deposits grow on top of the hot junctions. Assanis and Friedmann  found that the initial deposit formation over a surface causes the most significant reduction in time-averaged heat flux into the surface. In another study, Guralp et al.  tried to utilize this event to track the deposit growth by correlating the change in peak temperature phasing to the deposit thickness.
Photon and Radiance Based Measurements
Phosphor thermometry, or laser-induced phosphorescence (LIP), is a non-contact optical measurement technique. In this technique, an inorganic oxide or other phosphorescent material is applied onto the surface of interest using a non-fuorescent binder. When the material is excited with laser light, the resulting temperature-dependent phosphorescence is used to determine the surface temperatures by a temporal method or spectral intensity ratio method . The main advantages of the LIP technique are its higher temporal resolution and its insensitivity to scattered laser light due to long lifetime of phosphorescence. The LIP technique has been employed to measure the temperatures of piston crowns, intake valves and cylinder walls. Fuhrmann et al.  utilized phosphorescence decay technique viewing through a borescope and measured the metal engine cylinder wall temperatures to be about 60[degrees]C during motoring and 150[degrees]C during firing operation at the start of compression. Kashdan and others also employed the LIP technique to measure the piston temperature of an optically accessible engine [3, 19]. In another study, Knappe et al.  conducted phosphorescence decay measurements for the wall side and gas side of the phosphor dot, applied on a quartz liner. The authors concluded that LIP for in-cylinder measurements should be applied with caution and suggested a phosphor thickness of well below 30 [micro]n to minimize any insulating effect.
Infrared detection systems previously used include an IR microscope, IR camera, and spectral filters. Spikes  inserted small sapphire windows on a single-cylinder diesel engine and measured the temperatures that resulted on the faces of piston rings and lands using infrared emission during a firing cycle. A gold mirror was used to direct the infrared signal from the piston ring/liner contact to an upright IR microscope. In general, an IR camera has a much higher spatial resolution and increased sensitivity than the IR microscope and offers instantaneous capture of full field measurements .
Narrowband spectral filtered infrared imaging is another way of determining temperature of the surface of interest. Luo et al.  developed a dual wavelength infrared diagnostic for cylinder head surface temperature of an optical engine using an indium antimonide infrared detector. In that work sapphire emission was excluded from the infrared recordings by using two filters of center wavelength 3.275 [micro]m and 3.86 [micro]m, where sapphire was expected to have high transmissivity. Glass and like materials become opaque in the region of about 4-5 [micro]m . Hence, one way to determine surface temperature of a glass-like material is by measuring its emission using an IR filter at 5.1 [micro]m. However, the spectral responsivity of mid-wavelength infrared camera detectors significantly drops beyond 5 [micro]m wavelength . Recently, Buono et al.  installed an optical element, "AVL Visioknock lens", made from a sapphire lens in a modified spark plug. The IR spark plug sensor detected the intensity of radiation emitted from the hot piston surface. The radiation produced by fuel and the hot gases of combustion such as C[O.sub.2] and [H.sub.2]O dominated the IR signal. Hence, the IR spark plug was limited to taking recordings during intake and compression strokes before fuel injection.
In the current work, a technique is developed based on infrared radiation principles to measure the piston sapphire window temperatures of an optical engine. Using the proposed technique, piston window temperatures were measured during engine operation. In addition, soot formation was used as a variable to investigate the effect of piston temperature on combustion in the optical engine.
SPOT INFRARED-BASED TEMPERATURE (SIR-T) AND RADIANCE CALIBRATIONS
In this Spot Infrared-based Temperature (SIR-T) measurement technique, metal spots were vacuum-arc-deposited on sapphire window surfaces. Vacuum arc deposition is a physical vapor deposition technique in which an electric arc vaporizes and ionizes material from a cathode (metal in this case). The location and motion of the arc are controlled by an applied magnetic field. The vapor condenses and is deposited on the substrate (sapphire). Chromium was selected as the spot material as it bonded well to the sapphire surface. Chromium was vacuum-arc-deposited up to a spot thickness of approximately 200 nm at different locations on the piston window: outer top, bowl, and piston bottom as shown in Figure 1. The bowl spots and outer top spots are located at an axial distance t=15.5 mm and t=25 mm from the piston bottom surface as shown in Figure 2. An infrared camera was used to record the radiation along the metal spot, referred to as "radiance digital counts." All the results presented in this work were obtained when the window was positioned such that the piston bottom was facing the camera as shown in Figure 2. This window orientation was chosen to be consistent with the view of the combustion chamber looking through the piston window of an optical engine. The calibration procedure that correlates the radiance digital counts along the chromium spot image to the measured temperature is described in the following section.
Bench top oven heating experiments were used to calibrate the IR camera measurements versus piston top temperature measurements. The sapphire window with its bottom spots facing the camera was placed inside an oven. A time span of 1.5 hrs was required for the window temperature to reach 180[degrees]C. A K-type thermocouple was ceramic-bonded to the outer top surface and monitored using a temperature meter placed outside the oven. As seen in Figure 1, the chromium-coated sapphire piston window was placed in the oven with a black plate in the background.
A mid-wavelength infrared camera (FLIR Phoenix: MWIR Camera) was used whose viewable wavelength region is from 3-5 [micro]m. The camera incorporates an indium antimonide detector which is photovoltaic and generates current when exposed to infrared radiation. It is cryogenically cooled to 77 K using liquid nitrogen. The detector has 11[degrees] field of view and its response time is less than 1 ns, thus offering real-time measurement values. A Janos Technology ASIO 100 mm MWIR lens was attached to the detector. The integration time of the camera detector was set to 0.02 ms. A two-point non-uniformity correction (NUC) was performed before taking the IR recordings to avoid anomalies in recordings. Infrared images were recorded every 15-20[degrees]C during the time the window was air-cooled from 180[degrees]C to 75[degrees]C by turning off the oven heat and opening the oven door.
For a thermocouple reading, the corresponding averaged radiance digital counts were recorded along each region of interest (ROI) on the infrared image of chromium spots. After this step was repeated for several temperatures, a calibration curve was generated between the radiance digital counts of chromium spots and the surface temperature. Figure 3 shows an infrared image with radiance digital counts recorded at a window temperature of 180[degrees]C; also highlighted are the average count values along the spots on each of the three surfaces.
Using similar experimental and camera setups, radiance digital counts were calibrated for a chromium spot (approx. 200 nm thick) deposited on another sapphire piece having a thickness of 6.5 mm. The calibration data obtained for spots on the three surfaces of the window as well as for the spot on the sapphire piece at different temperature readings are shown in Figure 4. In this figure, the axial distance (t) from the piston bottom surface are noted along with the piston window locations. The dashed lines represent the fourth order polynomial fits. A common observation in Figure 3 and Figure 4 is the contrasting difference between the average counts of spots facing the camera (piston bottom) and those that are away from the image surface (outer top and bowl). Possible factors that could have given rise to these differences in digital counts among bowl and piston bottom spots were investigated further and will be presented in the next Section.
The temperatures measured using the SIR-T technique were compared to the thermocouple measurements as shown in Figure 5. The dashed line represents the identity line with respect to the thermocouple readings. The SIR-T measured temperatures were observed to deviate by a maximum of 10[degrees]C from the thermocouple measurements. Hence, the accuracy of the SIR-T technique is within 10[degrees]C for the considered working temperature range. The K-type thermocouple used in this work had a standard error above 2.2[degrees]C while the remaining error is attributed to data acquisition and curve fit calibration.
SURFACE-TO-SURFACE RADIANCE DIFFERENCES
The differences seen in radiance digital counts between chromium spots on the bowl and those on the piston bottom were better understood by identifying different sources of thermal emission as shown in Figure 6. The major thermal factors identified were 1) room air convection, 2) external radiance of background/oven walls, and 3) sapphire radiance.
The total radiance captured by the camera along a region of interest on the chromium spot can be written as,
[mathematical expression not reproducible] (1)
[mathematical expression not reproducible] (2)
where [tau] is the transmissivity in the IR spectrum. Subscripts "spot" and "wall" represent the chromium spot and wall behind the piston window. In the camera's spectral response of 3-5 [micro]m, there is negligible absorption by oxygen and nitrogen [25, 26]. Hence, absorptivity due to air is excluded in the above formulations.
The terms [I.sub.Spot_back] and [I.sub.Spot_front] indicate the radiance of chromium spots from its two sides: back side touching the sapphire surface and front side exposed to surroundings. Surface roughness has an effect on infrared emissivity . During vacuum-arc deposition of chromium onto the sapphire window, the surface roughness of the two sides of chromium spot could be different, thereby resulting in different emissivity. This possibility was checked by placing another spot-deposited sapphire piece behind the window. The front side of chromium deposited on the sapphire piece and the back side of chromium deposited on the sapphire window shared similar optical paths as shown in Figure 7. The corresponding radiance digital counts, shown in Figure 8, from back side of the chromium spot (on window) and front side of the chromium spot (on sapphire piece) differed by 6% to 10% depending on the temperature measured. A separate test was completed that confirmed the emissivity of front sides of chromium spots on the sapphire piece and piston bowl were similar such that comparisons between the two sides of the piston window could be made. Based on the ratio of radiance counts, the emissivity of front side of chromium spot is determined to be about 0.94 times the emissivity of back side of the chromium spot (see emissivity ratio curve in Figure 8).
When the oven door is opened to record calibration data, heat loss from the piston bottom surface due to room air convection leads to a temperature gradient across the piston window. To measure this effect, another thermocouple (not shown) was bonded to the piston bottom. After window heating and during the time the window was air cooled, both thermocouples showed temperatures within 1[degrees]C, showing a minimal temperature gradient between outer top and piston bottom surface. Thus, temperature gradients were assumed to have no effect on calibration data for the duration of calibration experiments.
Hydrocarbon fuel and combustion gases such as C[O.sub.2] have strong absorption bands in the camera's viewable wavelength region of 3-5 [micro]m [28, 29]. So, it is essential to evaluate the ability of the deposited chromium spots in blocking background radiation of unburned fuel, combustion gases, or the hot cylinder head while determining the window temperatures during engine firing. The blocking ability can be characterized by the transmissivity of the chromium spot ([[tau].sub.Spot]). Transmissivity of the chromium spot on the bowl was investigated by employing different background conditions as follows.
1. Configuration # 1 -No Plate - The window was heated without any plate placed in the background.
2. Configuration #2-Cold Plate - Placing a cold plate (at room temperature) in the background.
3. Configuration #3-Hotter Plate - The window was heated in the oven as in configurations #1 and #2. Another table top heater was used to heat the background plate separately to a temperature higher than the window temperature. This configuration provides useful information about the possible influence of background radiation during engine operation.
The measured temperatures from the calibration data for outer top and bowl spots are shown in Figure 9 with configurations #1 and #2. Here the region of interests 0 and 1 are along chromium spots on the bowl and the ROIs 2 and 3 are on the outer top. The spot temperature was approximately 113[degrees]C for both configurations. Hence, there was no influence on the radiance counts due to a colder background. Figure 10 shows the temperature measurements for bowl spots (0 and 1), outer top spot (2), and piston bottom spot (4) with configurations #1 and #3. In this figure, the bowl temperature ([approximately equal to]170[degrees]C) was within 1[degrees]C without a background plate and with a 260[degrees]C background plate. The piston bottom temperature was higher with a hotter plate background than without a hotter background plate. Overall, the piston bottom temperature measurements compare well with those of bowl and outer top, when the corresponding calibration curves are used. Based on these findings using different background media, it can be stated that the background does not seem to have a significant contribution to the radiance along the chromium spots. Hence, infrared transmissivity of the 200 nm thick chromium spots is very low and it was verified that the background radiance contributed very little to the chromium spot calibrations.
Transmissivity of the chromium spot ([[tau].sub.Spot]) was further investigated by painting one of the chromium spots on the bowl with matte-black paint (emissivity ~ 0.95). The main objective here was to identify any differences in the recorded radiance counts when two co-planar spots are exposed to different backgrounds. Figure 11 shows the window with one of the chromium spots painted on the bowl surface. The orientation of the window is the same, with the piston bottom surface being head-on with the camera as before. The window was then heated in the oven without background plate (configuration #1); Figure 12 shows the corresponding infrared image captured when the thermocouple reading of the heated window was 152[degrees]C. The ROI 0 represents the chromium spot whose opposite surface was painted and ROIs 1 and 2 represent the other two non-painted chromium spots on the bowl. The pixel-averaged temperatures were similar with or without the paint, confirming that the current chromium thickness of 200 nm is sufficient to make them optically thick in the camera's wavelength range. This test is useful to know that the chromium spots can block radiance from combustion deposits on the piston. These observations on transmissivity could be applied to chromium spots on the bowl as well as on the piston bottom surface. Therefore, from Equation 1, radiance (counts) captured by the camera along ROI on the chromium spot of the piston bottom surface represent the actual radiance emitted by only the chromium spot without any background influences.
[mathematical expression not reproducible] (3)
Sapphire and Chromium Radiance
On substituting the result [[tau].sub.Spot] = 0 in Equation 2, radiance of ROI along a chromium spot on the bowl can be re-written as,
[mathematical expression not reproducible] (4)
According to the steady state energy balance for radiative heat transfer (Kirchhoff's Law) as,
[mathematical expression not reproducible] (5)
The transmissivity of an 8 mm thick sapphire was measured to be greater than 80% in the wavelength range of 1-4 um and was on average approximately 65% in the range 4-5 [micro]m . In another study, reflectivity of a 5 mm thick sapphire specimen was measured to be 0.1 or smaller between 3.5-5 [micro]m . Under the current working temperatures and wavelength range of IR camera, the transmissivity of sapphire is expected to decrease with an increase in thickness and IR wavelength. Assuming the sapphire reflectivity to be significantly low, the emission of energy from sapphire should increase with thickness in accordance with Equation 5. In Figure 4, the ROI along the outer top spot (axial distance t=25 mm) showed highest radiance digital counts followed by ROI on bowl spot (t=15.5 mm), sapphire piece (t=6.5 mm), and on piston bottom spot (t=0 mm).
To summarize, the additional radiance contributed by the sapphire present between the chromium spot and the camera is the dominant source of differences seen in radiance digital counts between piston bottom spots and outer top/bowl spots. The slight difference in emissivity of front and back side of the chromium spot would have a secondary effect on the net radiance counts. The difference in emissivity between the two sides of the chromium spots does not affect the SIR-T temperature measurement, as the radiance calibration curves include the emissivity of the given spots.
PISTON TEMPERATURES DURING ENGINE OPERATION
The SIR-T technique was implemented to measure the piston temperature of a single-cylinder optical engine. Optical access to the pent-roof type combustion chamber was achieved using a sapphire piston insert and a 45-degree mirror at the bottom of a Bowditch extension. The intake system incorporates an electro-hydraulic actuated Multi Air[R] system to control valve timing and lift. The engine specifications are listed in Table 1.
During engine operation, the radiance digital counts along chromium spot ROIs on the sapphire window were recorded using the 45-degree mirror. A schematic of the experimental setup is shown in Figure 13. A gold-coated mirror was employed for IR imaging due to its high reflectivity (>97%) in the detector's spectral response range. It is to be noted that the mirror was not used while obtaining radiance calibrations. The acquisition speed of the current camera is limited to 60 frames per second (fps) with the required resolution, translating to one picture being taken every 150 CAD at 1500 rpm. This frame rate, therefore, provides only one usable picture per combustion event. The camera was externally triggered with a 5V transistor-transistor logic (TTL) pulse provided by the engine control system.
Recall that the temperature of a steady-state, warmed-up piston of a metal engine was in the range of 145-160[degrees]C . The temperature measurements in that work were accurate to within 10[degrees]C. In the current work, a target piston temperature of 140[degrees]C was chosen and combustion from natural gas port-fuel injection (PFI) was employed as a pre-heating source to attain this target. Natural gas, primarily composed of methane, has little tendency to produce soot because methane is a lower member in the paraffin family  and is expected to mix well. Therefore, natural gas combustion acted as an effective piston warm-up source with minimal window fouling. The engine operating conditions are listed in Table 2. The piston position convention is such that TDC represents compression top dead center. Natural gas at an injection pressure of 5.7 bar was port fuel injected at 10 CAD aTDC. The chosen spark timing offered stable combustion and the chosen fuel pulse width resulted in an indicated mean effective pressure (IMEP) greater than 6.0 bar. Natural gas injection started after the engine reached the set speed.
Figure 14 shows the SIR-T measured piston bowl and piston bottom temperatures of the optical engine recorded at the start of compression for the two engine speeds. The temperatures of bowl and piston bottom surfaces at the beginning of firing cycles were 85[degrees]C [+ or -]
5[degrees]C which is close to the coolant temperature. The bowl temperature raised to the target temperature of 140[degrees]C after approximately 400 natural gas PFI firing cycles. The bowl temperature during PFI warm-up recordings when repeated for a given operating condition at different barometric pressures were repeatable to within [+ or -]3[degrees]C. It should be noted that the cycle-by-cycle temperatures of optical piston did not reach a steady state value even after 400 cycles. However, no efforts were made to attain stabilized temperatures because of different heat transfer properties seen with a metal piston and an optical piston. Therefore, these warm-up cycles were employed to attain piston temperatures replicating metal engine steady temperatures, before carrying out gasoline direct injection (DI) combustion chamber visualizations.
TEMPERATURE UNDERESTIMATIONS DUE TO ENGINE IMPLEMENTATION
The calibration data were obtained for an isothermally heated piston window. The sapphire window thickness is an order of magnitude larger than that of the chromium spot. Thus, the heat produced during engine combustion at the bowl side takes more time to penetrate through the sapphire window thereby causing a temperature gradient from piston bowl to bottom surface. Conduction and convection heat losses would further add to the formation of the temperature gradient along the sapphire thickness. In contrast, the chromium spot heats up quickly along its small thickness given that the heat capacity of chromium is 448 J/kg-K compared to 763 J/kg-K for sapphire at room temperature. From Figure 14, it can be seen that the piston bottom temperature was 110[degrees]C when the bowl temperature was 140[degrees]C. The SIR-T measured bowl temperature was deduced from the radiance recorded which is realized to be a combination of radiance from the sapphire and the chromium spot. Hence, because the sapphire window is at a lower average temperature than during calibration, the SIR-T values obtained for the bowl spots are underestimated in order to balance the radiance recorded. Two possible piston temperature gradient cases are considered here to analyze the bowl temperature underestimations using the SIR-T technique. In the first case, the whole sapphire window was assumed to be at a constant temperature of 110[degrees]C. In the second case, the window was assumed to have a hypothetical temperature gradient that satisfies a linear radiance gradient.
Sapphire Window at 110[degrees]C
A worst case hypothetical scenario is considered wherein the piston bowl did not heat up to the target temperature of 140[degrees]C and instead isothermally warmed up to only 110[degrees]C. The radiance captured by the camera along the 'Bowl plane' spot is a combination of sapphire and chromium radiances. As the sapphire window was isothermally heated in the oven, a measure of radiance from sapphire can be obtained by subtracting the radiance digital count of a piston bottom spot (whose radiance is dominated by chromium) from that of a bowl spot. Sapphire radiance counts calculated for different temperatures are listed in Table 3.
In this worst case scenario, radiance digital count [C.sub.Sapphire @ 110[degrees]C] = 2490. The camera originally recorded a count, [C.sub.Bowl_Total]=6761 at the SIR-T measured temperature of 140[degrees]C. The required radiance count of a chromium spot on the bowl to compensate for the lower sapphire radiance is given as,
[mathematical expression not reproducible] (6)
By using the calibration curve of the bottom spot, this count would correspond to a temperature of approximately 170[degrees]C. Hence the chromium spot temperature and the window bowl surface would be underestimated by 30[degrees]C in this hypothetical scenario.
Temperature Gradient Along the Sapphire Window
During engine operation, the piston window would have a temperature gradient along its thickness unlike the uniformly heated window as in the calibration experiments As another hypothetical error analysis, the sapphire window was assumed to have a temperature gradient along its thickness, with the piston bottom temperature being 110[degrees]C. This particular temperature gradient is assumed to consist of temperatures that would result in a linear radiance gradient. Furthermore, in this linear radiance gradient, the average radiance across the gradient is assumed to be comparable to the average of the radiance at the two limits (bowl and piston bottom). Based on these assumptions, the radiance counts of chromium spots on the piston top can be formulated by re-writing Equation 6 as,
[mathematical expression not reproducible] (7)
Equation 7 was solved for T using an iterative process and it was found that a bowl surface temperature of 150[degrees]C satisfied this hypothetical temperature gradient. For this hypothetical temperature gradient, the SIR-T values for the bowl surface were underestimated by 10[degrees]C.
Based on these hypothetical error analyses, the piston window temperature measured using the current technique might be 10[degrees]C to 30[degrees]C (worst case) less than the actual value.
PISTON WARM-UP EFFECT ON SOOT FORMATION
Engine Operation: DISI Combustion
The requirements of the aforementioned piston warm-up strategy can be demonstrated by comparing combustion visualizations of an optical engine with and without the warm-up sequence. Table 4 lists the operating conditions of the optical engine. Again, the piston position convention is such that TDC represents compression top dead center. Natural gas was used for PFI injection and LEVIII E10 premium certification fuel was used for direct injection. The DI fuel was delivered at 200 bar pressure and the fuel pulse width was adjusted to attain stoichiometric air-fuel ratio as measured by an oxygen sensor installed in the exhaust manifold.
Combustion Imaging and Soot Detection
A Photron APX RS high-speed visible color camera was used to capture E10 combustion in the optical engine. Combustion events of 100 consecutive E10 DI firing cycles (not skip-fired) were captured viewing through the piston window and the 45-degree mirror at an imaging frame rate of 10k fps. This frame rate corresponds to a temporal resolution of about one frame for every 1.2 CADs at an engine speed of 2000 rpm. The camera was externally triggered with a 5V TTL pulse provided by the engine control system.
Natural luminosity of soot in the visible wavelengths has been shown to be orders of magnitude higher than the chemiluminescence of combustion gases, allowing for this property to be used to visualize in-cylinder soot formation processes . A thresholding approach was employed wherein the pixels of the color image whose red values exceeded intensity value of 200 were considered to be occupied by soot particulates. The pixels with intensities below this threshold were filtered to zero values. Figure 15 illustrates the performance of the pixel-thresholding algorithm in identifying fuel-rich or soot regions while filtering the remaining regions of the combustion image. The pre-mixed fame emitted a weak violet light which could be explained by the emission of hydrocarbon combustion intermediate CH* radicals in the wavelength range 400-450 nm .
Figure 16 shows the normalized soot index detected for 100 E10 DI firing cycles with and without the warm-up sequence. The normalized soot index is defined as,
[mathematical expression not reproducible] (8)
Here the image area was 512x512 pixels. Sample images of combustion for both engine operating modes are included here for illustration. From these soot index plots and sample images, the contrasting difference between the two operating modes is directly evident. Without the warm-up step, the normalized soot index was above 0.3% of the view area for the highest cycle. With warm-up, the maximum cycle-averaged soot was reduced by 70% and the 100 cycle average soot showed a reduction of 40% compared to without warm-up step. Based on these observations, it is inferred that the warm-up sequence aided in fuel vaporization and/or reduced wall impingement, resulting in less soot formation.
A Spot Infrared-based Temperature (SIR-T) technique was developed to measure the piston window temperature in a firing optical engine. Radiance values of chromium spots deposited on different surfaces of the sapphire window were calibrated from the thermocouple readings in a bench-top oven. The ability of the chromium spots to block radiance from combustion gases and wall deposits was tested using different background scenarios. The SIR-T temperatures of bowl surface and piston bottom surface temperatures matched well for a window heated in an oven. Accuracy of the proposed SIR-T technique is within 10[degrees]C for the working temperature range. Under engine operating conditions, the bowl temperature measured during two engine runs was repeatable to within [+ or -]3[degrees]C at engine speeds of 1500 rpm and 2000 rpm.
In addition, a piston warm-up strategy was employed for optical engine applications. Due to its low soot-forming tendency, natural gas PFI combustion was used as the warm-up heat source in this work. The increase in piston bowl temperature was recorded while operating the engine with natural gas PFI combustion. In this way, the required number of natural gas PFI cycles for the piston window to reach steady temperatures of a metal engine piston was determined. Because the calibration occurred in a uniformly heated oven, errors in the piston surface temperature measurements in a firing engine are likely. The piston surface temperature was 10[degrees]C less than the actual value for an assumed linear radiance gradient along the sapphire thickness while a worst case isothermal heating caused a temperature underestimation of 30[degrees]C. Such temperature underestimations majorly caused due to the additional sapphire radiation can be minimized by employing bandpass filters ranging from 3-4 [micro]m where the sapphire transmittance is greater than 0.85. In addition, the 45-degree mirror used for combustion visualizations could be included while doing calibration experiments.
The importance of employing such a warm-up sequence was demonstrated by noting soot formation during E10 DI combustion. Soot detected using its natural luminosity as a marker was observed to be significantly higher with a relatively colder piston than with the warmed-up piston and combustion chamber. Employing such a warm-up strategy for optical engines is critical in reaching the piston temperature values representative of a metal engine steady-state, warmed-up condition. The optical engine operation using the proposed temperature measurement technique is demonstrated to provide reliable experimental results.
[1.] Zhao, H., "Laser diagnostics and optical measurement techniques in internal combustion engines," 2012.
[2.] Kaan, M. and Kelestemur M.H., "Temperature and Stress Examinations on Pistons for Damage Prevention and Efficiency Enhancement: Carbon/Carbon Composite Piston vs Conventional Aluminum Pistons," in ASME 2015 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, American Society of Mechanical Engineers, 2015.
[3.] Kashdan, J. and Thirouard, B., "A Comparison of Combustion and Emissions Behaviour in Optical and Metal Single-Cylinder Diesel Engines," SAE Int. J. Engines 2(1):1857-1872, 2009, doi:10.4271/2009-01-1963.
[4.] Steeper, R. and Stevens, E., "Characterization of Combustion, Piston Temperatures, Fuel Sprays, and Fuel-Air Mixing in a DISI Optical Engine," SAE Technical Paper 2000-01-2900, 2000, doi:10.4271/2000-01-2900.
[5.] Spikes, H., "Infrared study of a piston-ring temperatures in a fired engine," Final report, Imperial Coll. of Science and Technology, London (UK), 1990.
[6.] Cho, K., Grover, R.O., Assanis, D., Filipi, Z. , "Combining instantaneous temperature measurements and CFD for analysis of fuel impingement on the DISI engine piston top," Journal of Engineering for Gas Turbines and Power, 132(7):072805, 2010.
[7.] Kopple, F., Seboldt, D., Jochmann, P., Hettinger, A. , "Experimental Investigation of Fuel Impingement and Spray-Cooling on the Piston of a GDI Engine via Instantaneous Surface Temperature Measurements," SAE Int. J. Engines 7(3):1178-1194, 2014, doi:10.4271/2014-01-1447.
[8.] Assanis, D. and Friedmann, F., "A Telemetry Linkage System for Piston Temperature Measurements in a Diesel Engine," SAE Technical Paper 910299, 1991, doi:10.4271/910299.
[9.] Hamzehei, M. and Rashidi, M., "Determination of piston and cylinder head temperature distribution in a 4-cylinder gasoline engine at actual process," in Proceedings of the 4th WSEAS Int. Conf. on heat transfer, thermal engineering and environment, Elounda, Greece. 2006.
[10.] Wing, R. and Saunders, O., "Oil film temperature and thickness measurements on the piston rings of a diesel engine," Proceedings of the Institution of Mechanical Engineers, 186(1):1-9, 1972.
[11.] Ladommatos, N., Xiao, Z., and Zhao, H., "The effect of piston bowl temperature on diesel exhaust emissions," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 219(3):371-388, 2005.
[12.] Kato, N., Moritsugu, M., Shimura, T., and Matsui, J., "Piston Temperature Measuring Technology Using Electromagnetic Induction," SAE Technical Paper 2001-01-2027, 2001, doi:10.4271/2001-01-2027.
[13.] Lee, H., Bang, H.W., Kauh, S.K., and Kim, S.I., "Piston-temperature measuring system using electromagnetic induction with verification by a telemetry system," International Journal of Automotive Technology, 12(3): 351-358, 2011.
[14.] Ebrahimi, K.M., Lewalski, A., Pezouvanis, A., Mason, B., "Piston Data Telemetry in Internal Combustion Engines," American Journal of Sensor Technology, 2(1):7-12, 2014.
[15.] Lee, H.W., Oh, C.H., Kauh, S.K., and Ha, K.P., "Development of a multipiston temperature telemetry system using Bluetooth networks," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 220(2):127-138, 2006.
[16.] Guralp, O., Hoffman, M., Assanis, D., Filipi, Z. , "Thermal Characterization of Combustion Chamber Deposits on the HCCI Engine Piston and Cylinder Head Using Instantaneous Temperature Measurements," SAE Technical Paper 2009-01-0668, 2009, doi:10.4271/2009-01-0668.
[17.] Omrane, A., Juhlin, G., Aldan, M., Josefsson, G. , "Demonstration of Two-Dimensional Temperature Characterization of Valves and Transparent Piston in a GDI Optical Engine," SAE Technical Paper 2004-01-0609, 2004, doi:10.4271/2004-01-0609.
[18.] Fuhrmann, N., Schneider, M., Ding, C-P., Brubach, J. , "Two-dimensional surface temperature diagnostics in a full-metal engine using thermographic phosphors," Measurement Science and Technology, 24(9):095203, 2013.
[19.] Kashdan, J. and Bruneaux, G., "Laser-Induced Phosphorescence Measurements of Combustion Chamber Surface Temperature on a Single-Cylinder Diesel Engine," SAE Technical Paper 2011-01-2049, 2011, doi:10.4271/2011-01-2049.
[20.] Knappe, C., Andersson, P., Algotsson, M., Richter, M. , "Laser-Induced Phosphorescence and the Impact of Phosphor Coating Thickness on Crank-Angle Resolved Cylinder Wall Temperatures," SAE Int. J. Engines 4(1):1689-1698, 2011, doi:10.4271/2011-01-1292.
[21.] Reddyhoff, T., Spikes, H., and Olver, A., "Improved infrared temperature mapping of elastohydrodynamic contacts," Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology, 223(8):1165-1177, 2009.
[22.] Luo, X., Yu, X., and Jansons, M., "Simultaneous In-Cylinder Surface Temperature Measurements with Thermocouple, Laser-induced Phosphorescence, and Dual Wavelength Infrared Diagnostic Techniques in an Optical Engine," SAE Technical Paper 2015-01-1658, 2015, doi:10.4271/2015-01-1658.
[23.] Vollmer, M. and Mollmann, K.-P., "Infrared thermal imaging: fundamentals, research and applications," John Wiley & Sons, ISBN-13: 978-3527407170, 2010.
[24.] Buono, D., Iarrobino, E., and Senatore, A., "Optical Piston Temperature Measurement in an Internal Combustion Engine," SAE Int. J. Engines 4(1):482-497, 2011, doi:10.4271/2011-01-0407.
[25.] Smith, K. and Newnham, D., "Near-infrared absorption spectroscopy of oxygen and nitrogen gas mixtures," Chemical physics letters, 308(1):1-6, 1999.
[26.] Faye, A., Kou, Q., Farrenq, R., and Guelachvili, G., "High-Resolution Fourier Transform Spectroscopy of 14 N 2: Analysis of the (1-0),(2-1) Bands of the B 3 [parallel]g-W 3 [DELTA]u System," Journal of molecular spectroscopy, 197(2):147-157, 1999.
[27.] Yu, H., Xu, G., Shen, X., Yan, X. , "Low infrared emissivity of polyurethane/Cu composite coatings," Applied Surface Science, 255(12):6077-6081, 2009.
[28.] Jansons, M., Lin, S., and Rhee, K., "Infrared spectral analysis of engine preflame emission," International Journal of Engine Research, 9(3):215-237, 2008.
[29.] Malkmus, W., "Infrared Emissivity of Carbon Dioxide (4.3-[micro] Band)," Journal of Optical Society of America, 53(8):951-960, 1963.
[30.] Oppenheim, U.P., and Even, U., "Infrared Properties of Sapphire at Elevated Temperatures," Journal of Optical Society of America, 52:1078_1-1079, 1962.
[31.] Lee, G.W., Jeon, S., Park, S-.N., Yoo, Y.S. , "Temperature and thickness dependence of IR optical properties of sapphire at moderate temperature," International Journal of Thermophysics, 32(7-8):1448-1456, 2011.
[32.] Papagiannakis, R. and Hountalas, D., "Combustion and exhaust emission characteristics of a dual fuel compression ignition engine operated with pilot diesel fuel and natural gas," Energy conversion and management, 45(18):2971-2987, 2004.
[33.] Van Basshuysen, R. and Schafer, F., "Internal combustion engine handbook-basics, components, systems and perspectives," 345, ISBN-13:978-0768011395, 2004.
[34.] Zizak, G., Flame emission spectroscopy: fundamentals and applications. 2000.
Ravi Teja Vedula
1497 Engineering Research Ct, Rm E104
Energy and Automotive Research Laboratory
East Lansing, MI-48824
The authors would like to express their gratitude to Lars Haubold for depositing the chromium spots at the MSU Fraunhofer institute, and Ed Timm for his assistance in bonding the thermocouple to the sapphire window.
C - Radiance digital counts
I - Infrared radiance
T - Temperature
[tau] - Transmissivity
t - Thickness
aTDC - After top dead center
bTDC - Before top dead center
CAD - Crank angle degree
CH* - Methylidyne radical
Cr - Chromium
DI - Direct injection
E10 - Gasoline fuel with 10% ethanol by volume
fps - Frames per second
IMEP - Indicated mean effective pressure
LIP - Laser-induced phosphorescence
MWIR - Mid-wavelength infrared
Ni - Nickel
NUC - Non-uniformity correction
PFI - Port-fuel injection
rpm - Revolutions per minute
SIR-T - Spot infrared-based temperature
TTL - Transistor-transistor logic
Ravi Teja Vedula, Thomas Stuecken, and Harold Schock
Michigan State University
Cody Squibb and Ken Hardman
FCA US LLC
Table 1. Engine Specifications. Combustion Chamber Pent-roof Stroke 90 mm Bore 84 mm Connecting Rod 152 mm Compression ratio 10.5:1 Table 2. Piston Warm-up Conditions. Engine Speed 1500 rpm, 2000 rpm IMEP ~ 6.0 bar Manifold Pressure 83-90 kPa (throttled) Port-Fuel Injection Fuel Natural Gas Fuel Injection Pressure 5.7 bar Start of Injection 10 CAD aTDC (Expansion stroke) Injection Pulse Width 9.0 ms Spark Timing 24 CAD bTDC Table 3. Radiance Digital Counts of Chromium Spots and Sapphire. Sapphire + Cr Cr T [C.sub.Bowl_Total] [C.sub.Bowl_Total] (in [degrees]C) (Pixel Count) (Pixel Count) 110 4700 2210 125 5640 2639 140 6761 3222 158 8445 3622 177 10537 4685 Sapphire T [C.sub.Bowl_Total] (in [degrees]C) [C.sub.Bowl_Total]) 110 2490 125 3001 140 3539 158 4823 177 5852 Table 4. Engine Operating Conditions. Engine Speed 2000 rpm IMEP 3.5 bar Manifold Pressure 65 kPa (throttled) Fuel Injection * Port-Fuel Injection Natural Gas (warm-up) * Direct Injection Ethanol Blended Gasoline [ElO] Injection Pressure * Port-Fuel Injection 5.7 bar * Direct Injection 200 bar IVO, IVC 270 CAD aTDC, 660 CAD aTDC PFI Start of Injection 10 CAD aTDC (Expansion stroke) DI Start of Injection 280 CAD bTDC (Intake stroke) Spark Timing 24 CAD bTDC