Comparison of Direct-Injection Spray Development of E10 Gasoline to a Single and Multi-Component E10 Gasoline Surrogate.
Adoption of direct-injection on engines has seen a rapid increase from very limited applications in 2007, to 38% in 2014, and 46% in 2015 [1, 2]. Fuel economy benefits have made GDI a competitive technology to achieve stringent Corporate Average Fuel Economy (CAFE) standards. However, higher particulate matter (PM) emissions from GDI engines needs to be addressed . Both fuel economy benefits and increased PM emissions are closely related to the fuel spray injection and the subsequent processes including fuel vaporization, fuel-air mixing and/or wall impingement. Thus, fuel spray characterization is essential to understanding and improving engine design and operation.
Due to the complex composition of commercial fuels (gasoline, diesel, kerosene, etc.) used in engines, fuel spray investigations have extensively used fuel surrogates to represent the commercial fuels, both for thermophysical and chemical characteristics . For example, in planar laser induced (exciplex) fluorescence (PLIF/PLIEF) techniques, it is often essential to use non-fluorescing fuel surrogates with 10% or less tracer dopants [5, 6, 7, 8], though qualitative investigations using commercial gasoline are possible [9, 10]. Commercial fuels contain aromatics, which fluoresce upon laser excitation. However, the aromatics have higher boiling points than other components, which may lead to differences in spatial distribution of aromatics and other components, thus making it difficult to represent early vaporization and mixing . The use of fuel surrogates and tracer dopants are the limitations of the PLIF/PLIEF spray diagnostics since the mixture may not be able to fully represent the spray injection of commercial fuels . The limitations are largely pertinent to how well fuel surrogates could represent the commercial fuel and what influence the addition of tracers has on the spray. It has been recognized that multi-component surrogates better represent key fuel properties including the distillation curve and chemical classes than single-component surrogates [L3,14].
Adoption of multi-component fuel surrogates is seen in multiple works. Scholz  used a near-standard colorless gasoline with bitumen content removed and without any tracers (fluorescing species are present in the fuel) in a fuel-air-ratio-LIF (FARLIF) application to measure the mixture equivalence ratios. Stevens  designed a six-component surrogate that ensured co-evaporation with three tracers representing the light, medium and heavy fraction in gasoline. This six-component surrogate was used by Williams  to investigate the effect of residual exhaust gas on quantitative PLIF. However, Steven's six-component surrogate did not use the volatility of a specific type of gasoline as the target. Styron  designed a four-component surrogate based on matching the distillation curve of the California Phase II gasoline. This surrogate could work with tracer pairs representing light and heavy fractions in gasoline, and were used in a PLIEF investigation of vapor and liquid distribution in a port fuel injection engine. Ma  formulated a five-component surrogate that both simulated gasoline's distillation curve and ensured co-evaporation with the same three tracers as in [16, 17]. These studies have generally focused on matching the surrogate's volatility to their targeted commercial fuels and co-evaporation of tracers with the surrogates. Validation studies that compare the spray injections of the fuel surrogate and the commercial fuels using optical diagnostics are seen in [20, 21]. However, the effects of adding tracers to fuel surrogates have not been fully understood from the previously mentioned works. In many works related to using PLIEF and PLIF it is an assumption that the use of fuel surrogate and tracer mixtures could reasonably represent the spray characteristics of the real fuel, without full proof to support the assumptions. Thus it is the objective of this study to understand and evaluate the effect of fuel surrogates and tracers on spray characteristics.
The experiments are carried out in an optically-accessible spray and combustion vessel which is 1 liter in volume and rated at 345 bar maximum pressure. Temperatures due to piston compression in an engine and the gas compositions are created by a lean/dilute preburn process. Temperatures below 473 K are achieved by electric heaters. A combustible mixture consisting of [C.sub.2][H.sub.2], [H.sub.2], [N.sub.2] and [O.sub.2] is ignited by a spark plug, and by adjusting the proportion of the combustible gas and [O.sub.2], different levels of [O.sub.2] are achieved to simulate different levels of exhaust gas recirculation (EGR). The injection event is triggered during the cool down after preburn is complete. Readers are referred to  for more details on the gas mixtures for preburn and the cool-down process. The vessel is equipped with six windows for optical access and experimental instrumentation, as well as eight corner ports for gas exchange valves and a pressure transducer. The spray is issued by a seven-hole Bosch HDEV-5 injector (S/N 0261500 147) for direct-injection applications, and the injector is mounted on one of the six windows on the side. The distribution of the seven nozzle holes is shown in Figure 1 by a front view image.
Shadowgraph and Mie scattering techniques are employed in the current study to visualize fuel sprays in the vapor phase and the liquid phase, respectively. A Z-type shadowgraph optical setup is used as shown in Figure 2(a). and the Mie scattering setup is shown in Figure 2(b). The Mie scattering setup used the same optical path as shadowgraph technique by replacing the shadowgraph light source with two different light sources while retaining all other optics. This setup ensures the same pixel size between images from two optical setups. A high speed camera FASTCAM SA 1.1 was used with a Nikon 85 mm lens. The frame speed was 20,000 frame per second, shutter duration was 14.3 us, and aperture size was f/1.4.
The experimental conditions used in the current study represent the injection at two timings during the operation of an engine: one condition is during the compression stroke and the other is near the top dead center (TDC). Details are shown in Table 1.
It is the objective of the current study to understand the effect of fuel and tracers on spray characteristics, thus fuel surrogates are formulated based upon the baseline fuel of CARB LEV III E10 Certification Fuel , referred to as "E10 Gasoline" hereafter. A summary of the fuel properties is shown in Table 2. The important properties for spray investigations are fuel density (important for spray penetration), latent heat of vaporization and distillation curve (both affecting vaporization and vapor-liquid equilibrium).
Use of single component fuel surrogates in spray investigations are often favored when used to study certain phenomena (e.g. diesel fuel ignition) and developing spray/chemical kinetic models. Important limitations of single-component surrogates include: inability to model multiple fuel properties and the detailed chemical processes, which could be overcome by using multi-component surrogates . Thus for this study, both single- and multi-component E10 Gasoline surrogates are used based upon the three important properties: fuel density, latent heat of vaporization and volatility.
In the current study, the interest is to understand the spatial distribution of E10 Gasoline's mid-range distillation in the fuel spray, thus n-heptane is selected based on the match of its boiling point to T50 of E10 Gasoline. As shown in Table 3, n-heptane has 8.3% lower density and 13.7% lower latent heat of vaporization compared to the E10 Gasoline.
For the multi-component surrogate, a palette of surrogate components is selected as: n-pentane, n-hexane, n-octane, n-decane and ethanol, whose properties spanned around those of E10 Gasoline and whose chemical classes are n-alkanes and alcohols. The reason to limit the chemical classes to be mainly n-alkanes is because the formulated surrogates need to work with PLIEF/PLIF tracers for a quantitative in-situ characterization of the spray mixture (which is the objective of a companion work, not the current one). Alcohols are added to replicate the latent heat of vaporization of the E10 Gasoline.
Reaction Design's Reaction Workbench code is used to calculate the fuel mixture's properties based on the target fuel property constraints (fuel density and distillation curve) and the selected surrogate component palette. Latent heat of vaporization is a constraint not implemented in Workbench and is calculated separately by mass-fraction-weighted latent heat of vaporization of single components. The final composition of the multi-component surrogate is given in Table 4 and is composed of: 20% ethanol, 4% n-pentane, 10% n-hexane, 36% n-octane and 30% n-decane by volume. This final composition has 10% more ethanol content compared to E10 Gasoline, and this done as a compromise between the density and the latent heat of vaporization of the multi-component E10 Gasoline surrogate. Among the five components, ethanol has the largest liquid density and the latent heat of vaporization compared to the rest of the compounds. As shown in Table 3, the formulated multi-component surrogate mixture has 3.5% lower liquid density and 15.6% higher latent heat of vaporization compared to the targeted E10 Gasoline. The distillation points at 10%, 50% and 90% volume fractions matches those of the E10 Gasoline within 3 K; however, these values are based on numerical models without experimental validations.
Also shown in Table 3 are the properties of fuel tracers, in which diethyl-methyl-amine (DEMA) and fluorobenzene are used as a pair of tracers for PLIEF technique, and 3-pentanone is used as a tracer for PLIF technique. Tracers are added to the multi-component surrogate and n-heptane with volume fraction of 9% DEMA, 2% fluorobenzene in the PLIEF mixture, and 10% 3-pentanone in the PLIF mixture. The fuel mixtures used in this work are summarized in Table 4. The multi-component surrogate is not used with DEMA and fluorobenzene since a companion work found that the liquid phase fluorescence intensity of this mixture is too low to characterize the sprays.
Image Processing Methods
High-speed shadowgraph and Mie scattering images are recorded and processed to obtain temporal development of the macroscopic spray characteristics Details on the processing method for shadowgraph and Mie scattering images are shown elsewhere : here, only a brief description is provided in terms of the workflow:
1. Background subtraction;
2. Thresholding grayscale images to binary images;
3. In Schlieren/shadowgraph images, pixel dilation and region closing operations are necessary to track the entire spray plume; such operation is not necessary for Mie scattering images;
4. Boundary tracking based on binary images;
A sample shadowgraph image is shown in Figure 3(a), and a sample Me scattering image is shown in Figure 3(b). Both images are shown with processed boundaries overlaid on the spray and definitions of key parameters. The injector is mounted horizontally pointing leftward, and the injector tip location is indicated by a yellow plus sign. By referencing to the SAE J2715 standard , the vapor/liquid penetrations are defined as the distance along the injector axis from the injector tip to the leading edge of the spray. In this study, the vapor and liquid penetrations are the distances along the x direction (indicated by the yellow dashed line) between the leading edges of the vapor and liquid envelopes and the location of the injector tip. Definitions of the spray angles and the spray bend angles are modified based on SAE J2715 standard to adapt to the spray in the current research: two lines are drawn at 2mm and 7mm downstream of the injector tip, and the spray boundaries between these two lines are linearly fitted. The spray angle is defined as the angle enclosed by the two linearly fitted lines and the spray bend angle is defined as the half of the absolute difference between the two angles. This definition is similar to SAE J2715 standard but the distances of the two lines are modified from 5 mm and 15 mm to 2 mm and 7 mm, particularly due to the images at high ambient temperature and pressure condition (TDC condition in Table 1) where a transition occurs along the spray edge that affects the linear fit result. The two conditions for calculating spray angles are shown in Figure 3(a) and Figure 3(b). The spray angle in Figure 3(a) is the sum of two angles shown by equation (1). while the spray angle in Figure 3(b) is the difference between two angles shown by equation (2). Under both conditions, the spray bend angles are both calculated as equation (3).
[[theta].sub.SprayAngle] = [[theta].sub.1] + [[theta].sub.2] (1)
[theta].sub.SprayAngle] = [[theta].sub.2] - [[theta].sub.1] (2)
[[theta].sub.SprayBendAngle] = ([[theta].sub.2] - [[theta].sub.1])/2 (3)
Sensitivities of the results to the thresholds in the image processing operations are shown in Figure 4. Results on the vapor and the liquid penetrations and the spray angles based on both image types are shown to be threshold insensitive to ensure proper interpretation of the results.
RESULTS AND DISCUSSIONS
Fuel Effects - E10 Gasoline & Surrogates
The penetration is an important macroscopic spray characteristic as it quantifies the fuel-air mixing process, which involves momentum exchanges between the fuel and the ambient charge gas. Comparisons of the vapor and the liquid penetrations of the three fuel mixtures (E10 Gasoline, the multi-component E10 surrogate and n-heptane) are shown in Figure 5 (A), (C) for the 452 K test condition, and in Figure 6 (A), (C) for the 680 K test condition. The penetration curves are averages of five individual injections. The injection durations are 1 ms in all cases. At the 452 K condition the spray reaches the extent of the field of view at 1.5 ms and the data reaches its maximum value.
To understand the statistical significance of the test repeats, results from the Welch ANOVA (analysis of variance) test  on the penetration curves are shown in Figure 5 (B), (D) and Figure 6 (B), (D). The ANOVA method allows evaluating the data to understand if the means of different groups are equal with assumptions of independence, normality and homogeneity of variances of the different groups of data. The null hypothesis states that all group means are equal. Based on the comparison of the calculated probability p with a designated significance level, the null hypothesis that the means from different groups of data are the same is retained or rejected. The Welch ANOVA method allows evaluating data with violations in the homogeneity of variances and is used in this study . From Figure 5 (B) and Figure 6 (B), there are two periods of time that correspond to the early phase of injection and the mid-portion of the spray development (0.05 - 0.30 ms & 0.70 - 1.00 ms for the 452 K condition, 0.05 - 0.10 ms & 0.75 - 1.45 ms for the 680 K condition) in which the p value is lower than the significance level of 0.05, which means the differences in the vapor penetrations shown in Figure 5 (A) and Figure 6 (A) during the two periods are statistically significant. The two periods during the early phase of injection do not occur in the liquid penetration curves as shown by Figure 5 (D) and Figure 6 (D); however, the differences in the liquid penetrations are statistically significant towards the end of injection for the two fuel mixtures under both test conditions.
With the previous statistical analysis, the results show that during the early phase of injection, n-heptane has larger penetration than the other two fuel mixtures. The spray velocity expressed by the Bernoulli's equation is:
[mathematical expression not reproducible] (4)
N-heptane has the largest Bernoulli velocity due to its lowest density among the three fuel mixtures, and this suggests that the early phase of the spray is velocity-controlled, and it is important to match the density of the fuel surrogate to the target fuel for this phase. The momentum of the spray can be expressed as:
[mathematical expression not reproducible] (5)
It is observed from equation (5) that the spray momentum remains the same regardless of the liquid mixture density. The fact that n-heptane spray has larger penetrations than the other two fuels suggest that the fuel air mixing among the three fuels are different.
Differences in the vapor penetration curves during the mid-portion of the spray development are statistically significant and occur under both test conditions as shown by Figure 5 (A) and Figure 6 (A), in which both fuels have larger penetrations than E10 Gasoline, with the largest penetrations from n-heptane. The hydraulic injection ends at approximately 1 ms after start of injection (ASOI), and the time durations cover the late half (452 K condition) and extend to after the end of the spray injection (680 K condition). This suggests that multi-plume GDI spray has transitioned away from being velocity-controlled during the early phase of injection. The differences are likely due to vaporization, mixing and plume-to-plume interactions.
Differences in the liquid penetration curves after 1.8 ms occur under both test conditions. Since the liquid front has recessed relative to its locations during the injection, the liquid penetration would be more properly described as the distance from the liquid front to the injector tip. N-heptane has the largest reduction in the distance from the liquid front to the injector tip and the multi-component E10 surrogate has the least reduction after the end of injection. The liquid front locations of n-heptane after 2.5 ms ASOI are almost zero under 680 K. This suggests liquid droplet vaporization occurs at the spray front. Since this time period is after the end of hydraulic injection, vaporization is believed to be the main reason for the differences. The larger distances from the liquid front to the injector tip of the multi-component E10 surrogate than E10 Gasoline is likely due to its larger latent heat of vaporization since their distillation points matches within 3 K. The smaller distance between the liquid front and the injector tip of n-heptane compared to E10 Gasoline is due to both its lower boiling point compared to the final boiling point of E10 Gasoline, as well as its lower latent heat of vaporization.
Spray angles and spray bend angles are also important macroscopic spray characteristics. In each test, spray angles and spray bend angles are characterized for each frame. A single value for the spray angle and the spray bend angle in each test are available from the quasi-steady state between 0.5 ms and 1.0 ms ASOI for tests with Mie scattering diagnostics, and for the 452 K condition with the shadowgraph diagnostics. The results are obtained by finding the mean and the standard deviation of all the data points within the quasi-steady state and reported in Table 5 and Table 6. For the 680 K test condition, no quasi-steady states are established in the shadowgraph images and the spray angle from each frame are plotted versus time ASOI in Figure 7.
Spray angle results are shown in Table 5. The vapor spray angles at 452 K condition are smaller from shadowgraph images than the liquid spray angles from Mie scattering images. The differences are not expected to be a result of threshold sensitivity of the spray angles since it was shown in Figure 4 that the results are insensitive to the thresholds. It is hypothesized to be due to the limitations of the shadowgraph technique with smaller density gradients at the periphery of the sprays. Within the same diagnostic, the differences are not statistically significant.
With an increase in the ambient temperature from 452 K to 680 K, reduced liquid spray angles and increased vapor spray angles are observed. The differences in the liquid spray angles are statistically significant between n-heptane and E10 Gasoline, but not between the multi-component E10 surrogate and E10 Gasoline. The smaller liquid spray angle of n-heptane is expected to be due to faster vaporization of n-heptane than E10 Gasoline and the multi-component E10 surrogate, primarily attributed to the lower latent heat of vaporization of n-heptane. The differences in the vapor spray angles are statistically significant during two time periods (0.6-0.8 ms and 1.8-2.05 ms shown in Figure 7). especially for the multi-component E10 surrogate compared to the other two fuels. The vapor spray angles under the 680 K condition are transient spray angles and the larger vapor spray angles of the multi-component E10 surrogate are unlikely to be explained by existing correlations on the spray angles which are based on quasi-steady state predictions.
Spray bend angle is a term that describes the actual angular deviation from the injector axis specifically for a single-plume spray, per SAE J2715 standard. Technically, multi-hole GDI injectors have a separate spray bend angle for each spray plume so that spray bend angle is not usually applied to multi-hole injectors. The attempt in the current study is to assess the bend of spray center axis relative to the injector axis. Results are shown in Table 6. The measured spray bend angles were similar under 452 K condition using both diagnostics and under 680 K using Mie scattering, with similar standard deviations. Those measured from the shadowgraph images differ from the rest and the measurements are with larger standard deviations. With these results the spray bend angle can be determined as 15[degrees] for this injector.
Spray characterization through the liquid and vapor penetrations, spray angles and spray bend angles provides important yet incomplete information regarding the sprays: transitions along the spray boundaries occur (see Figure 3) and they shift the spray from a cone shape so that spray angles no longer capture this characteristics; also, the radial expansion between the transition and the jet head is not characterized. As such, the spray shapes are compared by spatial distribution probability figures. Shown in Figure 8 are the temporal evolution of the spray spatial distribution probability figures by overlaying processed binary spray images from all five repeats under 452 K and 680 K ambient condition. Note that these figures are not direct overlays of the original Mie scattering and shadowgraph images. The image set include both Mie scattering and shadowgraph images, with three fuel mixtures, denoted as E10 Gasoline (E10), multi-component E10 surrogate (Multi), and n-heptane (Hep). Spray spatial distribution probability figures provide information on vapor/liquid penetrations, spray angles and spray bend angles previously discussed. They also reveal other important information:
1. Under both ambient temperature and pressure conditions, vaporization of n-heptane is faster than E10 Gasoline, while vaporization of multi-component E10 surrogate is similar to E10 Gasoline. This is shown by spray spatial distribution figures from Mie scattering images. Starting from 1.8 msASOI, the back-end of the liquid phase n-heptane spray has moved downstream relative to those of E10 Gasoline and the multi-component E10 surrogate, even though their spray front still align well (similar liquid penetration). By 2.4 ms ASOI, the area of the liquid spray region shrinks considerably under the 452 K condition, and almost disappear under the 680 K condition.
2. Under the 680 K ambient condition, faster vaporization of n-heptane alters the liquid phase spray shape compared to E10 Gasoline and the multi-component E10 surrogate, which is most obvious after the end of injection at 1.0 msASOI. Liquid phase spray shape of E10 Gasoline and multi-component E10 surrogate are similar under both 452 K and 680 K conditions.
3. Under the 680 K ambient condition, in addition to a decrease in the vapor/liquid penetration, there is also a decrease in the spray radial expansion.
4. Under the 680 K ambient condition, individual spray plumes appear less apparent compared to the 452 K ambient condition, in both liquid and vapor phases
5. Vapor phase sprays of all three fuel mixtures appear similar for a given ambient condition.
Effects on Fuel Properties
Tracers are blended with no more than 10% volume fraction in all fuel mixtures studied. Their influences on the fuel mixture density and the latent heat of vaporization are quantified by calculating volume fraction weighted density and mass fraction weighted latent heat of vaporization with and without the tracers.
The influence on the distillation curve is evaluated by calculating the distillation curves of the mixture with and without the tracer. The Workbench is not used here to evaluate the distillation curve since its fuel library does not contain the tracer properties. Instead, the process is approximated as a batch distillation vapor-liquid equilibrium (VLE) process by Peng-Robinson equation of state (EOS) , shown as equation (6). (7), (8), (9), (10).
P = [[RT]/[V - b]] - [[[alpha][alpha]]/[[V.sup.2] + 2bV - [b.sup.2]]] (6)
[alpha] = 0.45724 x [R.sup.2][T.sup.2.sub.c]/[P.sub.c] (7)
b = 0.07780 x [RT.sub.c]/[P.sub.c] (8)
[mathematical expression not reproducible] (9)
[kappa] = 0.37464 + 1.54226[omega] - 0.26992[[omega].sup.2] (10)
To avoid convergence problems associated with iterative solving of cubic EOS, the Peng-Robinson EOS is rewritten in the form of compressibility Z, for vapor and liquid phase respectively. Superscripts v is for vapor and l is for liquid. The equation sets when applied to a multi-component mixture, are shown as equation (11), (12), (13), (14), (15), (16), (17), (18) , with the subscripts i denoting each species in the mixture.
[mathematical expression not reproducible] (11)
[mathematical expression not reproducible] (12)
[mathematical expression not reproducible] (13)
[mathematical expression not reproducible] (14)
[[beta].sub.i] = [b.sub.i][P.sup.Sat.sub.i]/RT (15)
[mathematical expression not reproducible] (16)
[q.sub.i] [equivalent to] [[alpha].sub.i]/([b.sub.i]RT) (17)
[mathematical expression not reproducible] (18)
VLE calculation by EOS is obtained by equating the vapor and liquid phase fugacities of the same component in the mixture, both defined by equation (19) through (20).
ln[[phi].sub.i] = [Z.sub.i] - 1 - ln([Z.sub.i] - [[beta].sub.i]) - [q.sub.i][I.sub.i] (19)
[mathematical expression not reproducible] (20)
Application of equations (11), (12), (13), (14), (15), (16), (17), (18), (19), (20) to the mixture VLE problem in calculating the mixture bubble-point requires additional mixing rules shown by equation (21), (22), (23), (24), where subscript i denotes component property and the variables without subscripts denote mixture property.
[mathematical expression not reproducible] (21)
[beta] = bP/RT (22)
[mathematical expression not reproducible] (23)
[mathematical expression not reproducible] (24)
Estimation of the activity coefficient [[gamma].sub.i] in equation (23) is based on the UNIFAC (UNIQUAC Functional-group Activity Coefficients) method [33, 34]. Readers are referred to  for more details on the calculation method.
Results on property (density and latent heat of vaporization) comparisons are shown in Table 7. The addition of flow tracers has an overall negligible effect on the surrogate mixture property: the differences in liquid density relative to their base fuel are within 2%, and the differences in latent heat of vaporization are within 3% relative to the base fuel. For n-heptane, the effect of tracers on density and latent heat of vaporization are less with DEMA/fluorobenzene tracer pair than 3-pentanone. Distillation curves of these mixtures are shown by Figure 9. Note that the method for VLE problem-solving using Peng-Robinson EOS is different than the algorithm used in the Workbench code in which a Rachford Rice algorithm is used to evaluate the equilibrium ratio of component fraction in liquid to its fraction in vapor .
The addition of 3-pentanone to the multi-component E10 surrogate alters the distillation curve by increasing the evaporation temperature below 70% volume distilled, and decreasing the evaporation temperature above that point. The sharp turning point on the multi-component E10 surrogate distillation curve is smoothed by 3-pentanone. The overall root mean squared deviation (RMSD) between the two curves is 3.4 K. Addition of both 3-pentanone and DEMA/fluorobenzene pair alters the constant boiling point of heptane by decreasing the boiling point before 80% volume distilled and increasing the boiling point during the final stage of distillation. It is worth mentioning that the low boiling point of DEMA (338 K) compared to fluorobenzene (358 K) and 3-pentanone (373 K) is mainly responsible for the lower initial point boiling of the n-heptane/DEMA/fluorobenzene mixture compared to the n-heptane/3-pentanone mixture. Also, almost the entire distillation curve of the n-heptane/3-pentanone mixture is below 371 K. This is lower than the boiling points of both components, and it indicates that an azeotropic mixture is formed by n-heptane/3-pentanone. The overall RMSD between n-heptane/3-pentanone and n-heptane is 2.0 K, whereas the RMSD between n-heptane/DEMA/fluorobenzene and n-heptane is 3.3 K.
Effects on Spray Development
The effect of tracer(s) on fuel spray development is compared by the penetration curves from the fuel mixtures with and without the tracer(s) are shown in Figure 10. Generally the vapor and liquid penetrations of the fuel mixtures with tracers follow closely to that of the fuels without tracers before the EOI. After the EOI, slight deviations among the penetration curves from mixtures with or without tracers occur but do not form definitive trends. For example, for multi-component E10 surrogate and its mixture with 3-pentanone, the vapor penetration of the mixture with 3-pentanone is larger after the end of injection under 452 K condition, whereas it is smaller after the end of injection under 680 K condition. Statistical analyses using the Welch ANOVA test show that the differences in the penetrations (vapor and liquid) of the fuel surrogates with and without tracers are mostly statistically insignificant, only except for a few time points when the probabilities are less than the significance level of 0.05. In terms of the liquid penetrations towards the EOI, under 452 K, n-heptane/3-pentanone has larger liquid penetrations than both n-heptane and n-heptane/DEMA/fluorobenzene mixtures, and the Welch ANOVA test has confirmed that the differences are statistically significant. The larger mixture latent heat of vaporization contributes to the larger penetrations since the penetrations after 2 ms are after the EOI.
Spray angles of n-heptane, n-heptane/DEMA/fluorobenzene, n-heptane/3-pentanone are shown in Table 8. Under both test conditions, the vapor spray angles are statistically similar between n-heptane and both of its mixtures with PLIEF and PLIF tracers, as shown by the Welch ANOVA test results in Table 8 and Figure 10 (cB) and (dB). The liquid spray angles of the n-heptane/3-pentanone mixture under both conditions are larger than the other two fuel mixtures, which is hypothesized to be attributed to less vaporization due to its larger latent heat of vaporization as shown in Table 7. The Welch ANOVA test on the spray angles of n-heptane and n-heptane/DEMA/FB mixture shows that the p = 0.0718 (680 K) and p = 0.165 (452 K), and these are all higher than the significance level of 0.05, suggesting that the differences in the spray angles between n-heptane and n-heptane/DEMA/FB mixture has no statistical significance.
Spray angles of the multi-component E10 surrogate, and the multi-component E10 surrogate/3-pentanone are shown in Table 9 and Figure 12. The Welch ANOVA tests show that the spray angles of the two fuels are statistically similar, using both shadowgraph and Mie scattering techniques, and under both test conditions. Figure 12 show no quasi-steady state in terms of the spray angle.
Spray spatial distribution probability figures are shown in Figure 13 for n-heptane, multi-component E10 surrogate and their mixtures with tracer(s). Observations from this figure include:
1. Spatial distribution of the liquid and the vapor of multi-component E10 surrogate with and without tracer are similar to each other from the start of injection to late (2.4 ms ASOI) after the end of injection.
2. Vaporization of n-heptane/3-pentanone mixture spray appears to be slower than n-heptane and n-heptane/DEMA/fluorobenzene mixture, as evidenced by the liquid spatial distributions after 1.8ms ASOI under both 452 K and 680 K ambient conditions, which is related to the larger latent heat of vaporization by introducing 3-pentanone into the mixture. However, their vapor spatial distributions appear to be similar. This suggests that preferential vaporization may occur with this mixture.
Fuel Temperature Effects
When mounting the injector into the vessel, a cooling jacket is used to control the injector nozzle tip temperature by circulating coolant between the cooling jacket and an external chiller. In this study, the vessel is heated to 452 K. With cooling on, the injector nozzle tip temperature is brought down to 363 K.
Along with the tests conducted for this study, the effect of fuel temperature on fuel sprays are compared by turning on and off the cooling under one particular condition at 452 K and 4 bar ambient condition, with n-heptane/3-pentanone fuel mixture by shadowgraph images. The vapor penetration curves (average of five repeats) are shown in Figure 14. Differences in the vapor penetration curves are evident between 0.3-1.1 ms ASOI. The Welch ANOVA test results show that the differences in the penetration before 0.2 ms and between 0.3 and 0.9 ms have statistical significance. With the cooling turned off, the fuel temperature is higher, and the vapor penetration is consistently shorter. In terms of the vapor phase spray angle comparisons, the spray angle with cooling on is 36.7[degrees] with 3.3[degrees] as the standard deviation, the spray angle with cooling off is 34.3[degrees] and 3.5[degrees] as the standard deviation. The Welch ANOVA test performed on the two spray angles returns a p value of 0.00031, indicating that the difference is statistically significant. Both shorter vapor penetrations and smaller spray angles from a higher fuel temperature spray indicate structural differences as a result of the fuel temperature, and such differences are visualized in Figure 15 with a comparison of the vapor spray spatial distribution probability figure.
With the cooling turned off, a higher fuel temperature leads to faster fuel vaporization, which creates transitions along the spray edge that is commonly seen in sprays under 680 K condition with cooling turned on. The overall shapes of the spray also differ when the cooling is on than when the cooling is off. Instead of the spray spanning in different plumes under 452 K, the spray collapses to a single cloud of fuel which is similar to the shape observed at the 680 K condition. This result indicates that vaporization is affecting the mixing of the sprays from a multi-hole GDI injector, whether by a higher ambient temperature or by higher fuel temperature.
The use of laser diagnostics with flow tracers in fuel spray research is a common practice and often fuel surrogates are used rather than the complex commercial fuel mixtures. Two assumptions involved with these studies are: (1) fuel surrogates reasonably represent the targeted commercial fuels in terms of the spray characteristics; (2) addition of tracers does not alter the spray characteristics. It is the objective of the current study to investigate these two basic assumptions.
The fuel of interest is E10 Gasoline fuel. Two fuel surrogates are formulated, including n-heptane as a single component surrogate, and a multi-component E10 surrogate consisting of 20% ethanol, 4% n-pentane, 10% n-hexane, 36% n-octane and 30% n-decane by volume. PLIEF tracer pairs DEMA/fluorobenzene and PLIF tracer 3-pentanone are mixed with n-heptane and the multi-component E10 surrogate by volume fraction of 9% DEMA, 2% fluorobenzene in the PLIEF mixture, and 10% 3-pentanone in the PLIF mixture. Shadowgraph and Mie scattering techniques are employed to understand the temporal evolution of the macroscopic spray characteristics. The main observations in this study are outlined as follows.
Comparisons of E10 Gasoline, the multicomponent E10 Gasoline surrogate and n-heptane:
* Spray penetrations: statistically significant differences in the vapor and liquid penetrations occur during spray injection under both 452 K and 680 K conditions. The larger penetrations of n-heptane during the early phase of spray injection indicate that the sprays are velocity controlled. The differences towards the end of the hydraulic duration at 1 ms and after 2 ms indicate that differences exist in the vaporization and mixing of the spray.
* Spray angles: under the 452 K condition, both the liquid and the vapor spray angles of all three fuels mixtures are statistically similar. However, under the 680 K condition, differences among the three fuel mixtures occur, with the multi-component E10 surrogate closer to E10 Gasoline in the liquid phase during the quasi-steady state, and n-heptane closer to E10 Gasoline in the vapor phase during the transient state.
* The spray bend angle of the injector is 15 [degrees].
* Vaporization: n-heptane spray vaporizes faster than E10 Gasoline and the multi-component E10 surrogate, under both the 452 K and 680 K ambient condition, while the vaporization of the E10 Gasoline and the multi-component E10 surrogate are similar.
The effect of tracer addition to the fuel mixture:
* Fuel properties: the effect is small in the current study with the maximum difference between the properties being 3%.
* Spray penetrations: the penetrations of the surrogate mixtures with and without tracers generally follow each other well, with the exception that n-heptane/3-pentanone mixture has longer liquid penetrations than n-heptane and n-heptane/DEMA/fluorobenzene mixture after 2 ms in the liquid phase, which is hypothesized to be caused by differences in fuel vaporization.
* Spray angles: with slower vaporization of n-heptane/3-pentanone, its liquid spray angle is larger compared to the other two mixtures using n-heptane as the base fuel. The liquid spray angles of the multi-component E10 Gasoline mixture are similar to the mixture with 3-pentanone. Vapor spray angles of both n-heptane and the multi-component E10 surrogate with and without the tracers are statistically similar.
* Vaporization: slower vaporization of n-heptane/3-pentanone mixture is observed compared to n-heptane and n-heptane/DEMA/fluorobenzene mixture. The vaporization is similar between the multi-component E10 Gasoline surrogate and its mixture with 3-pentanone.
The effect of fuel temperature:
* The vapor penetrations and the vapor spray angles with cooling are larger than those without cooling. Structural differences appear in the vapor spray envelopes.
Based on these observations, it is concluded that:
* Single component surrogates could match the liquid penetrations prior to the end of injection under 680 K spray condition and the vapor/liquid spray angles under 452 K spray condition, and the general trend of the unmatched penetrations and spray angles could be replicated, except towards the end of injection where large differences in terms of vaporizations are observed
* Multi-component surrogates could match the spray characteristics in terms of the liquid penetrations and the liquid spray angles under both 452 K and 680 K condition, as well as the vapor spray angles under 452 K condition.
* Both 3-pentanone and DEMA/fluorobenzene have minor influences over the spray characteristics including the vapor/liquid penetrations/spray angles when mixed with both the multi-component surrogate and the single-component surrogate, however, differences in the liquid penetrations after the end of injection are evident due to differences in the vaporization.
Thus, in designing a fuel surrogate, it is important and recommended to match not only the density and the volatility but also the latent heat of vaporization to represent the spray characteristics, in terms of the vapor/liquid penetrations/spray angles, and its vaporization processes, as shown by comparisons of the multi-component E10 surrogate and the E10 Gasoline. It is also recommended in the selection of flow tracers to ensure similar volatility and latent heat of vaporization between the tracer and the base fuel to ensure similar vaporization between the fuel mixture with and without the tracers. Comparisons between the vapor and liquid envelopes of n-heptane and n-heptane/3-pentanone show that vapor envelopes are similar but liquid envelopes are different, and it is hypothesized that preferential vaporization occurs in the GDI spray of n-heptane/3-pentanone mixture under both test conditions. Further investigations by the use of fuel surrogates from this study into the local mixing will be done by the use of laser diagnostics (PLIEF/PLIF) to further understand the mixture quality during the spray injection process.
[1.] EPA, "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 Through 2015". 2015, U.S. Environmental Protection Agency: Rockville, MD.
[2.] Stacy Davis, S. W., Robert Boundy, Sheila Moore, "2015 Vehicle Technologies Market Report", Oak Ridge National Laboratory, 2015.
[3.] Khalek, I., Bougher, T., and Jetter, J., "Particle Emissions from a 2009 Gasoline Direct Injection Engine Using Different Commercially Available Fuels," SAE Int. J. Fuels Lubr. 3(2):623-637, 2010, doi:10.4271/2010-01-2117.
[4.] Tang, M., Zhao, L., Naber, J., Lee, S-Y. "Diesel Fuel Surrogate for Spray and Combustion Investigations.", Paper No. 1G11, presented at the 9th U.S. National Combustion Meeting, Cincinnati, May 17-20, 2015.
[5.] Froba, A. P., Rabenstein, F., Munch, K. U., et al.,"Mixture of triethylamine (TEA) and benzene as a new seeding material for the quantitative two-dimensional laser-induced exciplex fluorescence imaging of vapor and liquid fuel inside SI engines". Combustion and Flame, 1998. 112(1-2): p. 199-209. doi: 10.1016/S0010-2180(97)81768-3.
[6.] Todd, D. F., Michael, C. D., Boguslaw, G., et al., "Quantitative liquid and vapor distribution measurements in evaporating fuel sprays using laser-induced exciplex fluorescence". Measurement Science and Technology, 2009. 20(12): p. 125401. doi:10.1088/0957-0233/20/12/125401.
[7.] Yeom, J., Chung, S., Ha, J., et al., "A study on the behavior of evaporating diesel spray using LIEF measurement and kiva code". KSME International Journal, 2004. 18(12): p. 2310-2318. doi: 10.1007/BF02990236.
[8.] Melton, L. A., "Spectrally separated fluorescence emissions for diesel fuel droplets and vapor". Applied Optics, 1983. 22(14): p. 2224-2226. doi:10.1364/AO.22.002224.
[9.] Drake, M., Fansler, T., and French, D., "Crevice Flow and Combustion Visualization in a Direct-Injection Spark-Ignition Engine Using Laser Imaging Techniques," SAE Technical Paper 952454, 1995, doi:10.4271/952454.
[10.] Hentschel, W., Block, B., Hovestadt, T., Meyer, H. et al., "Optical Diagnostics and CFD-Simulations to Support the Combustion Process Development of the Volkswagen FSI[R] Direct-Injection Gasoline Engine," SAE Technical Paper 2001-01-3648, 2001, doi:10.4271/2001-01-3648.
[11.] Schulz, C. and Sick, V., "Tracer-LIF diagnostics: quantitative measurement of fuel concentration, temperature and fuel/air ratio in practical combustion systems". Progress in Energy and Combustion Science, 2005. 31(1): p. 75-121. doi:10.1016/j.pecs.2004.08.002.
[12.] Desantes, J. M., Pastor, J. V., Pastor, J. M., et al., "Limitations on the use of the planar laser induced exciplex fluorescence technique in diesel sprays". Fuel, 2005. 84(18): p. 2301-2315. doi:10.1016/j.fuel.2005.05.009.
[13.] Mueller, C. J., Cannella, W. J., Bays, J. T., et al., "Diesel Surrogate Fuels for Engine Testing and Chemical-Kinetic Modeling: Compositions and Properties". Energy & Fuels, 2016. 30(2): p. 1445-1461. doi:10.1021/acs.energyfuels.5b02879.
[14.] Mueller, C. J., Cannella, W. J., Bruno, T. J., et al., "Methodology for Formulating Diesel Surrogate Fuels with Accurate Compositional, Ignition-Quality, and Volatility Characteristics". Energy & Fuels, 2012. 26(6): p. 3284-3303. doi:10.1021/ef300303e.
[15.] Scholz, J., Wiersbinski, T., and Beushausen, V., "Planar Fuel-Air-Ratio-LIF with Gasoline for Dynamic Mixture-Formation Investigations," SAE Technical Paper 2007-01-0645,. 2007, doi:10.4271/2007-01-0645.
[16.] Stevens, R. E., Ma, H., Stone, C. R., et al., "On planar laser-induced fluorescence with multi-component fuel and tracer design for quantitative determination of fuel concentration in internal combustion engines". Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 2007. 221(6): p. 713-724. doi:10.1243/09544070jauto162.
[17.] Williams, B., Ewart, P., Wang, X., et al., "Quantitative planar laser-induced fluorescence imaging of multi-component fuel/air mixing in a firing gasoline-direct-injection engine: Effects of residual exhaust gas on quantitative PLIF". Combustion and Flame, 2010. 157(10): p. 1866-1878. doi:10.1016/i.combustflame.2010.06.004.
[18.] Styron, J., Kelly-Zion, P., Lee, C., Peters, J. et al., "Multicomponent Liquid and Vapor Fuel Distribution Measurements in the Cylinder of a Port-Injected, Spark-Ignition Engine," SAE Technical Paper 2000-01-0243, 2000, doi:10.4271/2000-01-0243.
[19.] Ma, X., He, X., Wang, J.-x., et al., "Co-evaporative multi-component fuel design for in-cylinder PLIF measurement and application in gasoline direct injection research". Applied Energy, 2011. 88(8): p. 2617-2627. doi:10.1016/j.apenergy.2011.01.043.
[20.] Yu, X., Luo, X., Jansons, M., Kim, D. et al., "A Fuel Surrogate Validation Approach Using a JP-8 Fueled Optically Accessible Compression Ignition Engine," SAE Int. J. Fuels Lubr. 8(1):119-134, 2015, doi:10.4271/2015-01-0906.
[21.] Zheng, Z., Lee, P., Shrestha, A., Badawy, T. et al., "Role of Volatility in the Development of JP-8 Surrogates for Diesel Engine Application," SAE Int. J. Fuels Lubr. 7(1):116-130, 2014, doi:10.4271/2014-01-1389.
[22.] Nesbitt, J. E., Johnson, S. E., Pickett, L. M., et al., "Minor Species Production from Lean Premixed Combustion and Their Impact on Autoignition of Diesel Surrogates". Energy & Fuels, 2011. 25(3): p. 926-936. doi:10.1021/ef101411f.
[23.] HaltermannSolutions, "Product Information for CARB LEV III E10 Certification Fuel Premium Octane (HF0862)". 2014.
[24.] Farrell, J., Cernansky, N., Dryer, F., Law, C. et al., "Development of an Experimental Database and Kinetic Models for Surrogate Diesel Fuels," SAE Technical Paper 2007-01-0201, 2007, doi:10.4271/2007-01-0201.
[25.] Chupka, G., Christensen, E., Fouts, L., Alleman, T. et al., "Heat of Vaporization Measurements for Ethanol Blends Up To 50 Volume Percent in Several Hydrocarbon Blendstocks and Implications for Knock in SI Engines," SAE Int. J. Fuels Lubr. 8(2):251-263, 2015, doi:10.4271/2015-01-0763.
[26.] Yaws, C. L., "Thermophysical Properties of Chemicals and Hydrocarbons". 2008, Norwich, NY, 13815: William Andrew Inc.
[27.] Acree W.E., Chickos J., J.S., "Phase Transition Enthalpy Measurements of Organic and Organometallic Compounds", in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, Mallard W.G. Eds. P.J. Linstrom, Editor. 2016, National Institute of Standards and Technology: Gaithersburg, MD, 20899.http://webbook.nist.gov, (retrieved June 2, 2016).
[28.] Tang, M., Zhao, L., Lee, S.-Y., and Naber, J., "Effect of Combustion on Diesel Spray Penetrations in Relation to Vaporizing, Non-Reacting Sprays," SAE Technical Paper 2016-01-2201, 2016 doi:10.4271/2016-01-2201.
[29.] SAE International Surface Vehicle Recommended Practice, "Gasoline Fuel Injector Spray Measurement and Characterization," SAE Standard J2715, Iss. Mar. 2007.
[30.] Trujillo-Ortiz, A. a. H.-W. "Welchanova: Welch ANOVA Test for Unequal Variances. [WWW document]. http://www.mathworks.com/matlabcentral/fileexchange/37121-welchanova". 2012.
[31.] Liu, H., "Comparing Welch's ANOVA, a Kruskal-Wallis test and traditional ANOVA in case of Heterogeneity of Variance", in Department of Biostatistics. 2015, Virginia Commonwealth University: Richmond, Virginia.
[32.] Peng, D.-Y and Robinson, D. B., "ANew Two-Constant Equation of State". Industrial & Engineering Chemistry Fundamentals, 1976. 15(1): p. 59-64. doi:10.1021/i160057a011.
[33.] Smith J.M., V. N. H. C., Abbott M.M., "Introduction to Chemical Engineering Thermodynamics". 7th Edition ed. 2005, New York: McGraw-Hill.
[34.] Fredenslund, A., Jones, R. L., and Prausnitz, J. M., "Group-contribution estimation of activity coefficients in nonideal liquid mixtures". AIChE Journal, 1975. 21(6): p. 1086-1099. doi:10.1002/aic.690210607.
[35.] Reaction Workbench 15131,"Reaction Design: San Diego". 2013.
The authors would like to acknowledge FCA US LLC for the support and collaborations on the current work.
GDI - Gasoline direct injection
LIEF - Laser-induced exciplex fluorescence
LIF - Laser-induced fluorescence
CAFE - Corporate average fuel economy
DI - Direct injection
PLIF - Planar laser-induced fluorescence
PLIEF - Planar laser-induced exciplex fluorescene
FARLIF - Fuel-air-ratio LIF
EGR - Exhaust gas recirculation
TDC - Top dead center
DEMA - Diethyl-methyl-amine
ASOI - After start of injection
EOI - End of injection
VLE - Vapor liquid equilibrium
EOS - Equation of state
UNIFAC - UNIQUAC Functional-group Activity Coefficient
RMSD - Root mean squared deviation
ANOVA - Analysis of variance
Meng Tang, Jiongxun Zhang, Xiucheng Zhu, Kyle Yeakle, Henry Schmidt, Seong-Young Lee, and Jeffrey Naber Michigan Technological University
FCA US LLC
Table 1. Test conditions. Condition Compression Injection pressure, bar 100 Injection duration, ms 1 (trigger signal length) Fuel temperature, K 363 Charge density, kg x [m.sup.3] 3.0 Charge pressure, bar 4 Temperature, K 452 Charge gas [N.sub.2] Condition TDC Injection pressure, bar Injection duration, ms Fuel temperature, K Charge density, kg x [m.sup.3] 9.9 Charge pressure, bar 20 Temperature, K 680 Charge gas 0% [O.sub.2] after preburn Table 2. Fuel properties for E10 Gasoline certification fuel . Carbon, wt% 82.52 Hydrogen, wt% 13.78 Oxygen, wt% 3.7 Ethanol, vol% 10 Sulfur, ppm wt 10 Density, kg/[m.sup.3] 744 Latent Heat of Vaporization, kJ/kg 417  10%: 332 Distillation, K 50%: 371 90%: 431 Net Heat of Combustion, MJ/kg 41.84 Table 3. Comparison of E10 Gasoline, multi-component surrogate, n-heptane, and fluorescent tracer properties, DEMA, fluorobenzene and 3-pentanone. Density, Latent Heat of Distillation kg/[m.sup.3] Vaporization, /Boiling Point, K   kJ/kg  10% 50% 90% E10 Gasoline 744 417 332 371 431 Multi-comp 718 482 329 372 431 Surrogate n-heptane 682 360 371 DEMA 702 365 338 Fluorobenzene 1019 359 358 3-pentanone 810 450 375 Table 4. Summary of the fuel mixtures used in this work. Baseline fuel E10 Gasoline Multi-component 20% ethanol, 4% n-pentane, 10% n-hexane, 36% E10 surrogate n-octane and 30% n-decane by volume Single-comp. E10 n-heptane surrogate PLIEF mixture 89% n-heptane, 9% DEMA, 2% fluorobenzene by vol. PLIF mixture 1 90% multi-component surrogate, 10% 3- pentanone by vol. PLIF mixture 2 90% n-heptane, 10% 3-pentanone by vol. Table 5. Spray angles for E10 Gasoline, the multi-component E10 surrogate and n-heptane sprays from both shadowgraph images and Mie scattering images under test conditions of 452 K and 680 K. Spray angle from Shadow. Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test E10 Gasoline 37.9 [degrees][+ or -] 3.2[degrees] 452 K, Multi-compo. 37.2[degrees] [+ or -] 4.2[degrees] p = 0.59 4 bar E10 surr. F = 0.54 n-heptane 37.8[degrees] [+ or -] 4.1[degrees] E10 Gasoline 680 K, Multi-compo. Figure 7 20 bar E10 surr. n-heptane Spray angle from Mie Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test E10 Gasoline 39.6[degrees] [+ or -]2.8[degrees] 452 K, Multi-compo. 39.6[degrees] [+ or -] 2.6[degrees] p= 1.00 4 bar E10 surr. F =0.0043 n-heptane 39.6[degrees] [+ or -] 2.8[degrees] E10 Gasoline 32.1[degrees] [+ or -] 4.0[degrees] 680 K, Multi-compo. 31.4[degrees] [+ or -] 4.9[degrees] p = 2.5e-6 20 bar E10 surr. F=14.8 n-heptane 28.3[degrees] [+ or -] 3.4[degrees] Table 6. Spray bend angles for E10 Gasoline, multi-component E10 surrogate and heptane sprays from both Shadowgraph images and Mie scattering images under test conditions of 452 K, 4 bar and 680 K, 20 bar. Spray bend angle Test condition & Fuel from Shadowgraph images/standard deviation 452 K, E10 Gasoline 15.6[degrees] [+ or -] 1.7[degrees] 4 bar Multi-compo. 15.6[degrees] [+ or -] 1.5[degrees] E10 surr. n-heptane 15.7[degrees] [+ or -] 1.4[degrees] E10 Gasoline 19.5[degrees] [+ or -] 4.6[degrees] 680 K, Multi-compo. 16.7[degrees] [+ or -] 8.0[degrees] E10 surr. 20 bar n-heptane 19.6[degrees] [+ or -] 4.5[degrees] Spray bend angle Test condition & Fuel from Mie scattering images/standard deviation 452 K, E10 Gasoline 15.3[degrees] [+ or -] 1.3[degrees] 4 bar Multi-compo. 15.1[degrees] [+ or -] 1.6[degrees] E10 surr. n-heptane 15.1[degrees] [+ or -] 1.2[degrees] E10 Gasoline 15.6[degrees] [+ or -] 1.8[degrees] 680 K, Multi-compo. 15.0[degrees] [+ or -] 1.9[degrees] E10 surr. 20 bar n-heptane 13.9 [degrees] [+ or -] 1.4[degrees] Table 7. Property comparisons of surrogates with and without tracers, percentage changes in parentheses are relative to their base surrogates. Latent Heat of Density, kg/[m.sup.3] Vaporization, kJ/kg n-heptane 682 360 n-heptane/ 695 (+2%) 371 (+3%) 3-pentanone n-heptane/DEMA/ 691 (+1%) 361 (+0%) fluorobenzene Multi-comp surrogate 718 482 Multi-comp surrogate/ 727 (+1%) 479 (-1%) 3-pentanone Table 8. Spray angles for n-heptane, n-heptane/3-pentanone and n-heptane/DEMA/fluorobenzene sprays from both shadowgraph images and Mie scattering images under test conditions of 452 K, 4 bar and 680 K, 20 bar. Spray angle from Shadow. Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test n-heptane 37.9[degrees] [+ or -] 4.1[degrees] 452 K, n-heptane/ 37.4[degrees] [+ or -] 2.8[degrees] p = 0.74 4 bar 3-pentanone F = 0.31 n-heptane/ 37.3[degrees] [+ or -] 3.4[degrees] DEMA/FB n-heptane 680 K, n-heptane/ Figure 11 20 bar 3-pentanone n-heptane /DEMA/FB Spray angle from Mie Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test n-heptane 39.6[degrees] [+ or -] 3.3[degrees] 452 K, n-heptane/ 40.5[degrees] [+ or -] 2.6[degrees] p=0.010 4 bar 3-pentanone F=4.8 n-heptane/ 38.8[degrees] [+ or -] 2.8[degrees] DEMA/FB n-heptane 28.3[degrees] [+ or -] 3.4[degrees] 680 K, n-heptane/ 30.2[degrees] [+ or -] 3.7[degrees] p=0.019 20 bar 3-pentanone F = 4.1 n-heptane/ 29.5[degrees] [+ or -] 3.6[degrees] DEMA/FB Table 9. Spray angles for the multi-component E10 surrogate and the multi-component E10 surrogate/3-pentanone sprays from both shadowgraph images and Mie scattering images under test conditions of 452 K, 4 bar and 680 K, 20 bar. Spray angle from Shadow. Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test 452 K, Multi-compo. 37.2[degrees] [+ or -] 4.2[degrees] p = 0.29 E10 surr. 4 bar Multi-compo. 37.9[degrees] [+ or -] 2.8[degrees] F= 1.1 E10 surr. /3-peutauoue Multi-compo. E10 surr. 680 K, Multi-compo. Figure 12 20 bar E10 surr. /3-pentanone Spray angle from Mie Test condition & Fuel images [+ or -] standard deviation, p & F value from Welch ANOVA test 452 K, Multi-compo. 39.6[degrees] [+ or -] 2.6[degrees] p = 0.38 E10 surr. 4 bar Multi-compo. 39.2[degrees] [+ or -] 2.8[degrees] F = 0.79 E10 surr. /3-peutauoue Multi-compo. 31.4[degrees] [+ or -] 4.9[degrees] E10 surr. 680 K, Multi-compo. 32.4[degrees] [+ or -] 3.8[degrees] p=0.20 20 bar E10 surr. F=1.7 /3-pentanone
|Printer friendly Cite/link Email Feedback|
|Author:||Tang, Meng; Zhang, Jiongxun; Zhu, Xiucheng; Yeakle, Kyle; Schmidt, Henry; Lee, Seong-Young; Naber, J|
|Publication:||SAE International Journal of Fuels and Lubricants|
|Date:||Jun 1, 2017|
|Previous Article:||Experimental Assessment of a Novel Instrument for the Injection Rate Measurement of Port Fuel Injectors in Realistic Operating Conditions.|
|Next Article:||Numerical Study on Flash Boiling Spray of Multi-Hole Injector.|