3.1.1 The formation of spray

Figure 3 shows a sample of images of the spray formation of JetA-1 at an injection pressure of 30psi and 80% threshold analysis. The threshold analysis consists of 255 levels (i.e., from 0 to 255) within a pixel. The range of the specific percentage of the threshold reduced noise in the captured images. The 80% threshold analysis indicates the threshold analysis was performed at 204 level that converted the colour image to a black-and-white.

Figure 3.

A typical captured spray Image by PIV data acquisition system of Jet A-1 at 30psi.

As soon as fuel was injected into the chamber, the fuel droplets began to disintegrate and form a mist at the end of the spray. Changes in droplet size were observed throughout the spray process. A reduction in droplet size can be found in three different regions after the start of injection (ASI). The first reduction in the droplet size was observed at ASI = 120µs, in which the diameter of the droplet reduces from 0.22 to 0.18µm. The size of the droplet decreased further in the secondary breakup region (ASI = 340µs), where the size of the droplet reduces from 0.16 to 0.10µm. Finally, the droplet was in the form of mist at ASI = 620µs.

The formation of Jet A-1 spray at various injection pressures is illustrated in Fig. 4. The spray droplet could be clearly seen at low injection pressures, whereas the formation of mist was more evident at higher injection pressures.

Figure 4.

Spray image by PIV data acquisition system of Jet A-1 at the tested injection pressures.

Figure 5 shows the spray formation comparison of each tested fuel measured at 120psi. This figure shows the dependency of spray formation on fuel properties, particularly the density and viscosity. BD100 had the lowest spray area and largest intact droplet size compared with the other tested fuels. The evaporation zone of BD100 was substantially longer than those of the other four samples. H100 injection produced smaller droplets and a broader spray zone. Similar patterns of spray development were observed for the J100 and blended samples.

Figure 5.

Five different tested fuel samples Spray image by PIV data acquisition system at 120psi.

The lengths of each breakup region are compared in Fig. 6. The presence of a breakup zone in the captured images was digitally improvised. The uncertainty in the interpretation of the optical measurements of these breakups was acknowledged in each instance. The majority of the quantitative information on the break-up zone has been collected from conductivity probe measurements made with fuel sprays and “break points” determined from penetration continuity. The length was obtained by measuring the distance between the two regions based on the diameter reduction discussed earlier. Generally, it was observed that the breakup lengths measured from the tip of the nozzle decreased with an increase in the injection pressure for all the tested fuels. At a particular injection pressure, BD100 had the highest breakup length, whereas Jet A-1 had the shortest breakup length compared with the other tested fuels. This breakup is due to the unstable flow of sinuous disruption [Reference Chen, Li and Zhang26]. A higher rate of disruption breaks the liquid sheet at half-wavelength intervals into ligament droplets for mist formation [Reference Sivakumar, Vankeswaram, Sakthikumar, Raghunandan, Hu and Sinha27].

Figure 6.

Breakup length at different injection pressure for primary breakup, secondary breakup and mist formation.

3.1.2 Spray cone angle

The formation of the spray cone angle is significantly affected by the ambient density [Reference Yu, Yin, Deng, Jia, Ye, Xu and Xu28]. However, in this study, the ambient density was kept constant to investigate the influence of injection pressure on the spray cone angle. The cone angle was measured by detecting the spray boundaries from the spray images and the measured cone angle for every pair of captured images from the start of the injection until the end of a cycle of complete spray acquisition [Reference Fan, Hashimoto, Nishida and Ozawa29]. It is worth mentioning that the spray cone angle plays a crucial role in determining the fuel-air mixing, combustion efficiency and emission levels during gas turbine operation. The relationship between the spray cone angle and the aforementioned parameters was not investigated in this study. The spray cone angle measured in this study was only used to evaluate the formation of the cone angle for a wide range of fuel properties.

Figure 7.

Influence of injection pressure with spray cone angle measured at the start of injection (ASI).

Figure 7 shows the variation in the spray cone angle recorded after the start of the injection (ASI) under various injection pressures. At each injection pressure, a fluctuation in the spray cone angle was observed during the early stages of the spray. However, the fluctuation became more stable after 1.2 milliseconds (ms). Once spray stability was achieved, a reduction in the spray cone angles was clearly observed as the injection pressure increased. Increasing the injection pressure was essential for evenly distributing the spray. At a low injection pressure, the spray area was dispersed more widely owing to the low velocity of the boundary particles than at higher injection pressures. The wide dispersion of the spray results in a larger spray cone angle. At a particular injection pressure, blending HEFA with Jet A-1 produced a lower spray cone angle than pure HEFA. The spray angle of the blend decreases as the percentage of HEFA in the blend decreases. The lowest spray cone angle was produced by the BD100.

The spray cone angles for each fuel are varied around 2–14% as injection pressure increased. The formation of the spray angle is more scattered at lower pressures. At higher pressures (P = 120psi), the spray angles of all fuels, except BD100, were closer to each other. The spray cone angle of BD100 decreased significantly as the pressure increased from 60 to 120psi. This reduction was due to the density and viscosity of BD100, which were higher than those of the other fuels tested. The fuel blends (J40H60 and J50H50) provided a maximum of 9% variation in the spray cone angle as the injection pressure increased from 30 to 60psi.

3.1.3 Spray tip penetration

Spray tip penetration is defined as the maximum distance of the spray after injection into stagnant or flowing air [Reference Lim, Razak and Mazlan30]. The spray penetration of jet fuel at various injection pressures is shown in Fig. 8. The spray penetration was measured at 0.2ms after the time start of injection. The correlation between fuel properties and penetration length can be observed by examining the total spray length at a particular injection pressure. The longest penetration was produced by BD100, whereas the shortest penetration was produced by Jet A-1. The longest penetration length of BD100 is associated with the higher density and surface tension of the fuel compared with other fuels. However, the surface tension did not affect the penetration length of H100. Although H100 had the lowest density and surface tension, the effect of viscosity on penetration length was highly significant. The same observation was made in Ref. (Reference Kumar, Karmakar, Kumar and Basu31), who observed a higher rate of particle disintegration and dispersion with a lower viscosity and density of fuel. In addition to fuel properties, turbulence, divergence, kinetic energy, particle collisions and aerodynamic forces contribute to the continual breakdown of fuel particles. Blending 60% of HEFA with 40% Jet A-1 provides closer penetration length with Jet A-1.

Figure 8.

Spray tip penetration of the tested fuels at a various injection pressure.

An increase in the injection pressure with spray length was clearly observed from 60 to 120psi. This trend is similar to those observed by Mohan et al. [Reference Mohan, Yang, Tay and Yu32] and Kurachi et al. [Reference Kurachi, Serizawa, Wada, Kato and Ito33]. However, the increase in the penetration length with injection pressure was not identical at low pressures. As the pressure increased, the fuel with high density propagated longer than the low-density fuel. Despite the density, the viscosity of the fuel also contributed to spray tip penetration. The viscosities of J40H60 and J50H50 were lower than that of BD100. A lower viscosity tends to decrease the friction between the fuel and nozzle, thereby increasing the exit velocity. Similarly, J40H60 and J50H50 had lower densities than pure biodiesel, which enhanced the exit velocity. Droplets with lower viscosities are prone to deformation and breakup in later stages and produce significant momentum loss, restricting the propagation of the spray. The inertia of the fuel increases as the density increases, making it more resistant to the environment, thus allowing longer spray tip penetration. However, the effects of viscosity and density act accordingly in all zones during the spray development.

Figure 9.

Evaporation time of the tested fuel at different injection pressure.

3.1.4 Droplet lifetime

Droplet lifetime is defined as the time required by a liquid particle to fully vaporise and evaporate. In a gas turbine engine, fuel is injected using an atomiser to produce fine droplets. The droplets evaporate and become vapour before mixing with air in the combustion chamber. Atomisation is required to increase the liquid area, thereby enhancing evaporation. The droplet lifetime defines the nature of the atomisation of the fuel; the fuel has a lower lifespan to evaporate proportionally to clean combustion. However, droplet lifetime plays a crucial role in evaporation and combustion. Figure 9 shows the evaporation times of the fuels at different injection pressures. The evaporation time was measured experimentally by recording the time taken by the particle after the start of injection until it diminished. The droplet lifetime decreased as the injection pressure increased from 30 to 120psi. Injecting the spray at 120psi resulted in the shortest evaporation time after the injection started. The longest injection time is observed at the lowest injection pressure. Injecting fuel at high pressure increases the droplet velocity. Because the fuel was injected into stagnant airflow, a large velocity difference was created between the air and fuel spray. The large velocity difference enhanced the aerodynamic effects and developed turbulence, consequently promoting the disintegration of the fuel droplets. At a particular injection pressure, the BD100 droplet exhibited the longest lifetime, whereas the lowest droplet lifetime was observed for HEFA. A longer droplet lifetime was observed as the percentage of HEFA in the HEFA/Jet A-1 mixture decreased. The lifetime of the J50H50 droplet was closer to that of the Jet A-1 droplet at both lower and higher injection pressures. Under similar ambient conditions, fuels with lower viscosities and surface tensions exhibit rapid droplet disintegration and dispersion. Notably, in the present study, the droplet lifetime was measured under ambient conditions. The droplet lifetime of the fuel during the actual operation of a gas turbine engine is different. Droplet lifetime in a gas turbine engine is influenced by factors such as residence time, variation in droplet size, fuel properties, combustion conditions, turbulence and mixing, burner design and combustion system.

3.2.1 Mean droplet diameter (D10)

The mean droplet diameter (D10) is the average of all droplet diameters and provides information on the droplet size in the fuel spray sample. The D10 values of the tested fuels at various injection pressures are shown in Fig. 10 and were measured in millimeters. The mean diameter was calculated using the arithmetic mean diameter (D10) as shown in Equation (1): This equation was proposed by Ref. (Reference Kumar, Karmakar, Kumar and Basu31).

(1) \begin{equation}{D_{10}} = \frac{{\mathop \sum \nolimits_{i = 1}^k {D_i}}}{{\mathop \sum \nolimits_{i = 1}^n {N_i}}}\end{equation}

Figure 10.

Droplet size (mm) distribution in axial direction at different injection pressure.

In Equation (1), ${N_i}$ is the number of droplets with ${D_i}$ diameter. In general, D10 is used to simplify the data analysis [Reference Mayhew and Wood34]. As shown in Fig. 10, the average droplet size decreased consistently as the injection pressure increased, except for J50H50, which exhibited a slight increase at higher injection pressures. As expected, the D10 of BD100 is the highest amongst the fuels, which explains the longest droplet lifetime of the fuel and shows a similar pattern to that of Fan et al. [Reference Fan, Hashimoto, Nishida and Ozawa29]. The difference in droplet size is associated with the kinematic viscosity and surface tension of the droplets throughout the spray [Reference Lee and Reitz35].

3.2.2 Droplet concentration

Figure 11 illustrates the total number of average droplets counts in the spray region at various injection pressures by processing a captured image of a fully developed spray. Image processing in PIVlab software detected almost all droplets at an 80% threshold level. The number of droplets increased progressively as the injection pressure increased from 30 to 120psi. At lower injection pressures, the total number of droplets produced by HEFA, J50H50 and J40H60 was comparable to that of Jet A-1, which was approximately 930 droplets. As the injection pressure increased, Jet A-1 produced the highest number of droplets (approximately 10,614 droplets) compared with the other tested fuels. In contrast, the total number of droplets produced by BD100 was lower than those produced by the other fuels. The increase in the droplet number is also slower than that of other fuels, mainly owing to their viscous properties. According to this result, the spray droplet concentration is mainly determined by the continuous breakup of the spray droplets until mist formation and detection of all atomised droplets in the fuel spray region.

Figure 11.

Number of droplets detected for each pressure.

3.2.3 Droplet velocity profile

Figure 12 depicts the integral velocities of all the tested fuels and their profiles along the spray region at an injection pressure of 120psi. The integral velocity of the fuels was obtained directly from the PIV. PIV measures the particle velocity through a combination of the time interval and displacement of the particles between the two images. This result is consistent with those reported by Naber et al. [Reference Naber and Siebers36] and Sieber [Reference Siebers37]. Among the tested fuels, biodiesel produced the highest droplet velocity when the injection pressure was increased. Increases in the droplet velocity increase the Reynolds number of the droplet, causing the droplet to travel farther at higher injection pressures. The droplets travelled from the tip in a downward direction and the individual velocity of the droplet decreased continuously. Consequently, the deformation of the fuel droplets will be smaller under the inertial force near the nozzle exit, in which the aerodynamic drag acting on the fuel droplets will be smaller. The velocity profile was calculated by comparative image sequence analysis using PIVlab software. Both HEFA and J40H60 showed similar results. Figure 12 shows the overall droplet velocity magnitude for the horizontal (u) and vertical (v) components along the spray length measured from the nozzle tip. The fluctuation in particle velocity is due to the formation of a vortex at a particular location. After a fuel droplet is injected into the combustion chamber, its motion is influenced by the turbulent vortex [Reference Du, Al-Rashed, Barzegar Gerdroodbary, Moradi, Shahsavar and Talebizadehsardari38]. The turbulent vortex caused the droplet to move into a non-smooth curve instead of a straight line [Reference Yin, E, Ding and Li39]. The development and quality of the spray structure are influenced by the formation and induction of vortices created by the velocity difference between the droplet and the ambient air. Soon after the formation, the wave shapes produced in the early injection stage disappeared, and the spray patterns began to recover uniform shapes. However, increasing the pressure minimised the fluctuation in the spray tip velocity.

Figure 12.

Integral of velocity magnitude from the tip of the nozzle as a function of axial direction at 120psi.