Comparative experimental study on macroscopic spray characteristics of various oxygenated diesel fuels

Under high ambient pressure (5 MPa) and different injection pressures (90, 120, and 150 MPa), a high‐speed imaging technique was carried out to comparatively investigate the macroscopic spray characteristics of diesel with three different types of blended fuel in a constant volume chamber. The oxygenated fuels were n‐butanol (B), pine oil (P), and 2,5‐dimethylfuran (DMF). All their blending ratio with diesel were 20%. Results showed that less viscosity could be improved the spray characteristics of the fuel in the range of experimental conditions. Then, the tested fuels had a longer penetration and a greater spray area with increasing the injection pressure from 90 to 150 MPa. On the other hand, the percentage increases in the mean spray cone angle of D100, B20, P20, and DMF20 were 3%, 4.4%, 2.4%, and 2.9%, respectively. At the same experimental condition, the spray penetrations of DMF20 and P20 were larger than that of D100, but the spray penetration of B20 was basically similar to D100. Besides, the performance of the spray cone angle and spray area were D100 < B20 < P20 < DMF20. In addition, the comprehensive influence was that blending oxygenated fuels would be a benefit for developing fuel atomization and the air–fuel mixture of conventional diesel fuel.

sources, cetane number, energy density, and mutual solubility with diesel fuel. [7][8][9] Much of the oxygenated fuel work that has been published to date has covered the combustion and emissions performance of diesel engines. 10 Valentino et al. 11 found that with the increase in the n-butanol blend ratio and oxygen content, the cetane number within the cylinder decreased, and the ignition delay could be prolonged, thereby improving the combustion process. Huang et al. 12 pointed out that enriching a certain proportion of n-butanol in diesel fuel reduced soot and carbon monoxide (CO) emissions. The experimental investigations by Vallinayagam et al. 13,14 showed that the calorific value of pine oil was close to that of diesel fuel. Pine oil/diesel blended fuel (vol. 50%) could provide comparable engine performance with pure diesel at full load, and it allowed CO, unburned hydrocarbon, and soot reductions with a slight nitrogen oxide (NO x ) penalty rate. The research on diesel/ 2,5-dimethylfuran (DMF) fueled engines 15 indicated that blending 2,5-dimethylfuran could decrease soot emissions effectively. When the blending ratio was 40%, the near-zero soot emission of the engine could be obtained. To further understand the role of 2,5dimethylfuran in reducing soot emission, Liu et al. 16 found that the lower cetane number of 2,5dimethylfuran increased ignition delay, which was the main factor for soot reduction.
The combustion mode of diesel engines can be classified into two categories: space diffusion combustion of the in-cylinder mixture and wall-film combustion of the liquid fuel, of which the former is the primary mode. 17 Therefore, the injection and atomization of the fuel within the cylinder play an important role in the quality of the combustion process of diesel engines. 18 However, there are few efforts and coherent information to date concerning the spray characteristics of the aforementioned oxygenated diesel fuels. To get a further understanding of the effect of oxygenated fuels on the macroscopic spray process, it is necessary to carry out relevant visualization research and make a comparison of the spray characteristics of different oxygenated diesel fuels, and therefore, to provide a certain reference for the exploration of the optimal fuel ratio of diesel engines in the future.
In summary, the author collected the spray patterns of different oxygenated diesel fuels in a constant volume combustion bomb through high-speed photograph technology and compared them with pure diesel fuel. Then, it could be revealed that the influence of different oxygenated diesel fuels (n-butanol, pine oil, and 2,5dimethylfuran) on macroscopic spray characteristics of the engine.

| Preparation of oxygenated fuel
In the experiment, n-butanol (C 4 H 9 OH), pine oil (C 10 H 18 O + C 10 H 16 ), and 2,5-dimethylfuran (C 6 H 8 O) were blended with commercial diesel (#0 Sinopec, China V) according to the volume ratio of 1:4. Three kinds of blended fuels (B20, P20, and DMF20) were prepared, which represented that the blended fuel containing 20% n-butanol and 80% diesel, 20% pine oil and 80% diesel, and 20% 2,5-dimethylfuran and 80% diesel, respectively. Besides, pure diesel (D100) was used as the reference fuel. The main property parameters of the tested fuels at 20°C are listed in Table 1, 12,16 in which the property parameters of pine oil are derived from the sample inspection report of the Guangzhou Institute of Energy Testing. 8 The surface tension of the fuel was obtained according to the empirical formula (1). 19 Where ρ represents density, g/cm 3 . σ denotes surface tension, N/m.

| Experimental setup and procedures
The schematic diagram of the macroscopic spray visualization test bench is depicted in Figure 1. The experimental test bench mainly consists of a constant volume combustion chamber, a high-pressure common rail fuel injection system, and a high-speed camera acquisition system. The three sides of the self-designed constant volume combustion chamber were equipped with optical windows, and inlet and exhaust valves were mounted on the other side of the chamber. The fuel injector, thermocouple, and pressure transducer were arranged on the detachable head cover. All joint parts of the constant volume chamber are equipped with specially machined sealing gaskets, ensuring that the T A B L E 1 Main property parameters of the tested fuels at 20°C. Abbreviations: B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel.

Molecule
chamber maintains an acceptable sealing performance under high ambient pressure. A BOSCH CP2 highpressure common rail fuel injection system was used in this work, and the highest injection pressure can be reached at 220 MPa. A BOSCH solenoid valve single-hole injector was positioned centrally on the head cover of the chamber. Two halogen tungsten lamps placed in front of the quartz window on both sides of the chamber were employed to illuminate the fuel spray. A high-speed charge-coupled device camera, Photron FASTCAM-SA7, and a Tokia macro lens were used to record the spray development patterns, as shown in Figure 2A.
The test process was carried out under no light and normal temperature (25°C) conditions. At first, highpressure nitrogen gas was charged to provide the ambient pressure required by the experimental study. Then, the chamber pressure was captured and monitored in realtime by using a 6052C pressure sensor and charge amplifier manufactured by KISTLER. In the experiment, the fuel injection parameters were set by a high-pressure common rail console, and then the motor and the highpressure pump started to work. When the injector injected the fuel, the electronic control unit sent a TTL5V trigger signal to the high-speed camera through F I G U R E 1 Schematic diagram of the macroscopic spray visualization test bench. (1) Fuel tank, (2) light source, (3) air compressor, (4) pressure transmitter, (5) high-pressure nitrogen gas, (6) electronic control unit, (7) program counter, (8) common rail, (9) high-pressure pump, (10) electromotor, (11) high-speed charge-coupled device (CCD) camera, (12) fuel injector, and (13) constant volume combustion chamber.
F I G U R E 2 Schematic of the used high-speed camera and synchronic control mode. (A) High-speed camera and (B) synchronic control mode. the filter plate. As shown in Figure 2B, there is an adjustable dwell (ahead time of trigger) between injector injection signal and camera trigger signal, which is mainly due to the requirement of background removal in the subsequent image processing. The electronic control unit first triggers the shutter signal of the high-speed camera so that the first image acquired is a background image without the spray, and then sends the injector injection signal for fuel injection. The high-speed camera and synchronization acquisition of the spray pattern were implemented with the assistance of the halogen tungsten lamp of the background lighting. The main condition parameters of the experimental measurement are summarized in Table 2. Figure 3 summarizes the image processing process of spray macroparameters. The spray images captured in the experiment were preprocessed by the functions of Matlab software, such as image enhancement, morphological operation, and so on. It was necessary to compare and analyze the renderings and determine the optimal preprocessing algorithm. Then, the image segmentation and edge detection algorithms were used to find the favorable binary segmentation map. In this regard, the required spray macroparameters for the experimental analysis, such as spray cone angle and penetration distance, could be obtained by using the written functions. To ensure the accuracy of experimental measurement, each group was sprayed three times under the same test condition, and then the average value was calculated to get the corresponding results. For repeatable experimental procedures, the transformation of the tested fuel was realized through the drain valve at the bottom of the fuel tank. After replacing the tested fuel, it was necessary to start up the high-pressure common rail injection system to clean the oil circuit. Then, the highpressure nitrogen was used for scavenging, and the exhaust valve was tightened after scavenging. Finally, nitrogen was charged into the constant volume bomb for the next group of spray tests. The duration between the two groups of tested fuels was more than 20 min. The measurement resolutions of the experimental equipment are summarized in Table 3. The definition of spray cone angle and tip penetration is plotted in Figure 4 with reference to the previous studies. [20][21][22] The spray tip penetration refers to the axial distance from the nozzle outlet of the injector to the front of the spray. Since there T A B L E 2 Summary of engine specifications. is no uniform definition for the spray cone angle to date, the present work adopts the method of Naber and Siebers, 23 and the intersection angle formed by the two tangent lines drawn from the nozzle outlet along the spray edge to the half of the penetration is used as the spray cone angle. 24 3 | RESULTS AND DISCUSSION

| Comparison of macroscopic spray patterns
To facilitate the comparison of the spray development of different oxygenated fuels, all images were programmed to the same exposure time and frame speed for each group of experiments. The timing method of after the start of injection (ASOI) was used for results analysis. 25 A total of 20 images were captured from the beginning of the spray at the nozzle outlet to the end of the injector. The ASOI of the first image was 0.1 ms, the time interval between subsequent adjacent images was 0.1 ms, and the ASOI of the last image was 0.2 ms. After the Matlab program processing, Figure   F I G U R E 5 Spray patterns of different oxygenated diesel fuels. B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel.
The distribution of the fuel concentration field can be seen intuitively in Figure 6. The brightness and darkness of the spray show the difference in the concentration of the fuel field. The brighter the spray image is, the greater the fuel concentration in the spray field, and the darker the spray image is, the closer the fuel concentration is to the ambient medium concentration. As shown in Figure 5, the concentration of the main spray area is higher than that of the spray edge, which indicates that the contour edge of the spray entrains the air and atomization evaporation faster. 26 With the further development of the fuel spray, the range of the high concentration in the main spray area is expanding. Variation of the oxygenated fuel does not affect the trend. In addition, by comparing the partial spray process of these four fuels, it is found that the main spray areas of B20, P20, and DMF20 are notably dense, and the edge contours of these oxygenated fuels are smooth. However, the flow field concentration of pure diesel is relatively diluted. Due to the large shear force on the spray front, the annular flow on both sides of the cone is obvious, and the interaction between partial fuel droplets and the environmental medium can be presented as well. 27 The reason for this phenomenon may be due to the fact that the higher viscosity of diesel (Table 1) affects the homogeneous and dispersion of fuel atomization, resulting in lower regularity when interacting with environmental gases. Thus, the radial diffusion velocity of pure diesel is increased compared with other oxygenated blended fuels.

| Effect of fuel property on macroscopic spray characteristics
Through the postprocessing procedure mentioned above, the results of the spray image were quantitively addressed, and the dot-line diagrams shown in Figures 6-8 were obtained. That is, the evolutions of the spray tip penetration, spray cone angle, and spray area as a function of the ASOI under the conditions of the injection pressure of 150 Mpa, the ambient pressure of 5 MPa, and the temperature of 25°C. It can be seen from the whole spray process in Figure 6 that whether it is diesel or the three oxygenated fuels, their spray tip penetrations show a similar linear growth trend with the increase of the ASOI. Then the growth rate of the penetration decreases slightly as the spray is stabilized. Furthermore, the spray tip penetration of DMF20 and P20 is greater than that of D100, and their average increment is 4.7 and 1.5 mm, respectively. However, there is no significant difference between B20 and D100 in terms of spray tip penetration. Figure 7 shows the evolutions of the spray cone angles of different oxygenated diesel fuels as a function of the ASOI. As illustrated in Figure 7, the variation trends of the spray cone angle of these four fuels are roughly the same. Their peak values are all presented at the initial stage of spraying and then gradually decrease and remain in a relatively stable value range. The variation of the spray cone angle in the whole spray process is small, which is in the range of 2°-3°f or each tested fuel. The average spray cone angle of these four fuels is D100 < B20 < P20 < DMF20 from small to large under the same injection and ambient pressure. In addition, the spray area is an intuitive reflection of the fuel F I G U R E 6 Spray tip penetrations of different oxygenated diesel fuels. ASOI, after the start of injection; B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel.
F I G U R E 7 Spray cone angles of different oxygenated diesel fuels. ASOI, after the start of injection; B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel. atomization quality. From the analysis of Figure 8, it can be observed that compared with the pure diesel at the same ASOI moment, the average spray areas of three oxygenated fuels have increased to some extent. Specifically, B20, P20, and DMF20 are increased by 1.4%, 3.8%, and 10.5%, respectively.
To clarify the intrinsic mechanism of various fuel atomization under the same test condition, the fuel properties such as viscosity, surface tension, and density should be considered. 28 As typical dimensionless parameters, Reynolds number and Weber number are usually used to characterize the breakup and atomization of different fuel droplets. On the basis of the general research method of spray stability, 29 two different spray modes, namely, Rayleigh mode and Taylor mode, can be distinguished according to the Weber dimensionless number (We) and the liquid-gas density ratio (Q). When the We/Q ≪ 1, it is defined as the Rayleigh mode. The surface tension plays a dominant role in this mode. The spray velocity and ambient density are relatively small, and the influence of the ambient gas on the spray can be ignored. When the We/Q ≫ 1, it can be defined as the Taylor mode. The inertial force prevails in this mode, and the interaction between gas and liquid cannot be neglected due to more significant spray velocity and ambient density. As listed in Table 4, the We is much larger than the Q, and thus the Taylor mode is to be investigated in the current work. In Taylor mode, the larger the Reynolds number and Weber number are, the better the atomization quality of the droplet. 30 As shown in Table 4, the relationship between Reynolds number and Weber number of the four fuels is inconsistent. This is because the fuel flows into the common-rail cavity after being pressurized by a highpressure pump and then injected into the cylinder through the injector. The instantaneous pressure and temperature of the fuel increase sharply, which results in an obvious variation of the fuel property. Relevant studies 31,32 demonstrated that when the fuel temperature exceeded 50°C, except for the kinematic viscosity, the surface tension, sound velocity, and elastic modulus all performed an increasing trend to varying degrees with the increase of pressure, and the normal state could no longer be used as the calculation medium. Thus, the Reynolds number, which characterizes the ratio of liquid inertial force to the viscous force, has a great influence on the spray development of the fuel. The order of the Reynolds number from high to low is DMF20 > P20 > B20 > D100. The reason is that the kinematic viscosity of the three oxygenated fuels is smaller than that of pure diesel (Table 1), thereby leading to a larger Reynolds number. Especially for DMF20, the kinematic viscosity is reduced by 32.8% compared with D100, and the density difference between these two fuels is only 0.013 g/cm 3 , which indicates the viscous force of fuel has a significant influence on the spray characteristics during the spatial development of the spray. Besides, the droplets with small viscous force are not easy to stick together, and the initial flow velocity of the spray is relatively high accordingly. The entrainment effect between the spray edge and the surrounding gas can be enhanced and improve the breakup and atomization of the fuel spray. 33 Therefore, the atomization mixing quality of different fuels presents D100 < B20 < P20 < DMF20, which is in good agreement with the experimental results of spray characteristics described above. Figure 9 plots the overall average spray cone angle (0.1-2 ms ASOI) of the four fuels under varying injection pressures when the nozzle diameter is 0.17 mm, the ambient pressure is 5 MPa, and the ambient temperature is 25°C. It can be observed from Figure 9 that with the F I G U R E 8 Spray areas of different oxygenated diesel fuels. ASOI, after the start of injection; B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel. Abbreviations: B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel.

| Effect of injection pressure on macroscopic spray characteristics
increase of injection pressure, the spray cone angles of the four fuels increase to some extent. When the injection pressure increased from 90 to 150 MPa, the average spray cone angles of D100, B20, P20, and DMF20 increased by 3%, 4.4%, 2.4%, and 2.9%, respectively. The reasons for this are as follows: First, the injection pressure has a direct effect on the fuel flow within the nozzle hole. According to the investigation of Marchi et al., 34 the air bubbles that enter the nozzle hole from the outlet and adhere to the wall could affect the spray cone angle, and the adhesion position of these air bubbles determined the actual spray cone angle to a certain extent. With the increment of injection pressure, the internal thrust of the spray was enhanced, and the adhesion bubble moved down. This indicates the fuel has a higher radial velocity and initial kinetic energy. Second, increasing the injection pressure leads to a smaller droplet diameter, making the entrainment effect between the droplets and the surrounding gas at the edge of the spray more intense, which is manifested as the largerscale and more numbered annular flow on both sides of the spray cones, and forms a larger spray cone angle consequently. In addition, the spray cone angles of B20, P20, and DMF20 are larger than that of D100 under three varying injection pressures of 90, 120, and 150 MPa. The phenomenon can be explained by the fact that the Reynolds numbers of these three oxygenated fuels are larger than that of pure diesel, and the loss of flow resistance in the nozzle hole is small. As a result, the jet velocity required for cavitation is low, and the radial turbulent pulsation velocity is relatively high, thus increasing the spray cone angle. However, the difference between them is small, and the maximum discrepancy is less than 1°. Figure 10 plots the spray tip penetration and spray area of different oxygenated diesel fuels at the end of the spray (2 ms ASOI) under three injection pressures of 90, 120, and 150 MPa. It is found that no matter what kind of tested fuel, the final spray tip penetration and final spray area present an increasing trend with the increment of fuel injection pressure. This phenomenon can be explained by the fact that when the injection pressure increases, the spray has a higher injection velocity to penetrate forward, which can also be confirmed by the spray front speed of D100 and P20 (Figure 11). After the fuel is ejected from the nozzle hole, the kinetic energy of the fuel is bound to increase, which requires a longer F I G U R E 9 Average spray cone angles of different oxygenated diesel fuels. B20, 20% n-butanol and 80% diesel; D100, pure diesel; DMF20, 20% 2,5-dimethylfuran and 80% diesel; P20, 20% pine oil and 80% diesel.
F I G U R E 11 Spray front speeds of different oxygenated diesel fuels. ASOI, after the start of injection; D100, pure diesel; P20, 20% pine oil and 80% diesel. axial distance to conduct sufficient energy exchange with the medium inside the constant volume combustion chamber, thereby extending the forward penetration. Moreover, elevating the injection pressure increases the pressure difference between the inside and outside of the nozzle hole. The increase in the inertial force of the droplet leads to a larger Reynolds number, which accelerates the breakup and atomization of the spray and aid in shortening the fuel-air mixing process. 35 Therefore, the spray area of higher injection pressure presents larger. As shown in Figure 11, compared with D100 at the same ASOI moment, the variation in the spray front speed of P20 is relatively stable, while the fluctuation in the spray front speed of D100 is pronounced under the influence of injection pressure.

| CONCLUSIONS
The macroscopic spray characteristics of different oxygenated diesel fuels, including diesel/n-butanol blends, diesel/pine oil blends, and diesel/2,5-dimethylfuran blends, have been carefully compared and analyzed.
Three injection pressures of various tested fuels are investigated in this study to clarify the consistency of the conclusions. The conclusions summarized from the experimental measurement are as follows: (1) Compared with pure diesel, the macroscopic spray patterns of three oxygenated fuels have a dense main spray area, larger high concentration range, and smooth edge contour. The spray development trend of the four tested fuels is consistent. (2) For different physical properties of tested fuels, based on the Taylor model, the viscosity force plays a dominant role in the macroscopic spray characteristics of various oxygenated diesel fuels.
(3) Under the tested experimental conditions, the spray tip penetration and spray area of oxygenated fuels increase with the increment of injection pressure, while the spray cone angle is less affected by increasing injection pressure, which is basically the same as that of pure diesel. (4) The spray tip penetrations of DMF20 and P20 are greater than that of D100, while the difference between B20 and D100 is negligible in terms of spray tip penetration. The order of the spray cone angle and spray area for four tested fuels from low to high is D100 < B20 < P20 < DMF20. This indicates that blending a certain amount of n-butanol, pine oil, and 2,5-dimethylfuran can effectively improve the atomization mixing process of diesel fuel.