Deep learning-based image processing for analyzing combustion behavior of gel fuel droplets

Agarwal, Vandan; Chitkariya, Puneet; Miglani, Ankur; Nandagopalan, Purushothaman; John, Jerin; Kankar, Pavan Kumar

doi:10.1016/b978-0-323-90789-7.00011-7

Cited by 3 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The HED algorithm-based image processing captures all the four parameters accurately namely, the area of rupture sites, centroid of rupture sites, the spatial movement of rupture sites, and the percentage overlap between successive rupture areas. The projected-area diameter is determined using an in-house developed deep learning-based holistically nested edge detection (HED) algorithm, as reported previously [ 33 , 34 , 35 , 36 ]. HED enables automated learning of multiscale and multilevel features in a droplet image and reconstructs a continuous droplet.…”

Section: Methodsmentioning

confidence: 99%

“…This is carried out by feeding an array of computed radial distances and their corresponding angles to the standard polar plot function from the matplotlib python library. Based on the edge detection, the maximum error in determining the droplet diameter is 3% [ 33 ]. The temporal evolution of the droplet oscillation frequencies is determined via the continuous wavelet spectra (CWT) of the time series signal of droplet diameter fluctuations in-built into the MATLAB R2021b version software.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Time-Varying Oscillatory Response of Burning Gel Fuel Droplets

Sharma

Miglani

2023

Gels

View full text Add to dashboard Cite

Gel fuel droplets exhibit disruptive burning due to the rupture of their gellant shell, which causes the release of unreacted fuel vapors from the droplet interior to the flame in the form of jets. In addition to pure vaporization, this jetting allows convective transport for fuel vapors, which accelerates gas-phase mixing and is known to improve droplet burn rates. Using high-magnification and high-speed imaging, this study found that the viscoelastic gellant shell at the droplet surface evolves during the droplet’s lifetime, which causes the droplet to burst at different frequencies, thereby triggering a time-varying oscillatory jetting. In particular, the continuous wavelet spectra of the droplet diameter fluctuations show that the droplet bursting exhibits a nonmonotonic (hump-shaped) trend, where the bursting frequency first increases and then decreases to a point where the droplet stops oscillating. The changes in the shell structure are captured by tracking the temporal variation of the area of rupture sites, spatial movement of their centroid, and the degree of overlap between the rupture areas of successive cycles. During the initial period (immediately following its formation) when the shell is newly formed, it is weak and flexible, which causes it to burst at increasingly high frequencies. This is because the area at and around the rupture site becomes progressively weaker with each rupture in an already weak shell. This is shown by a high degree of overlap between the areas of successive ruptures. On the other hand, the shell flexibility during the initial period is demonstrated by a reversal in the motion of rupture site centroids. However, at later stages when the droplet has undergone multiple ruptures, the depletion of the fuel vapor causes accumulation of gellant on the shell, thus causing the shell to become strong and rigid. This thick, strong, and rigid shell suppresses droplet oscillations. Overall, this study provides a mechanistic understanding of how the gellant shell evolves during the combustion of a gel fuel droplet and causes the droplet to burst at different frequencies. This understanding can be used to devise gel fuel compositions that result in gellant shells with tailored properties, and therefore, control the jetting frequencies to tune droplet burn rates.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Time-Varying Oscillatory Response of Burning Gel Fuel Droplets

Sharma

Miglani

2023

Gels

View full text Add to dashboard Cite

show abstract

“…This section comprises the details regarding the flame-scale imaging of both sets of fuels, i.e., the ethanol-based HPMC-3% and MC-9% gel fuels. Deep learning is applied for flame tracking and calculations of the area and perimeter of the flame [ 37 , 38 , 39 ]. The Schlieren imaging setup produces a flame of a discontinuous periphery.…”

Section: Methodsmentioning

confidence: 99%

Jetting Dynamics of Burning Gel Fuel Droplets

Sharma

Miglani

John

et al. 2022

Gels

View full text Add to dashboard Cite

Jetting in burning gel fuel droplets is an important process which, in addition to pure vaporization, enables the convective transport of unreacted fuel vapors from the droplet interior to the flame envelope. This aids in accelerating the fuel efflux and enhancing the mixing of the gas phase, which improves the droplet burn rates. In this study, Schlieren imaging was used to characterize different jetting dynamics that govern the combustion behavior of organic-gellant-laden ethanol gel fuel droplets. To initiate jetting, the gellant shell of the burning gel fuel droplet was subjected to either oscillatory bursting or isolated bursting, or both. However, irrespective of the jetting mode, the jets interacted with the flame envelope in one of three possible ways. Based on the velocity and the degree to which a jet disrupts the flame envelope, it is classified as either a flame distortion, a fire ball outside the flame or a pin hole jet (localized flame extinction), where the pin hole jets have the highest velocity (1000–1550 mm/s), while the flame distortion events have the lowest velocity (500–870 mm/s). Subsequently, the relative number of the three types of jetting events during the droplet lifetime was analyzed as a function of the type of organic gellant. It was demonstrated that the combustion behavior of gel fuels (hydroxypropyl methylcellulose: HPMC at 3 wt.%) that tend to form thin-weak-flexible shells is dominated by low-velocity flame distortion events, while the gel fuels (methylcellulose: MC at 9 wt.%) that facilitate the formation of thick-strong-rigid shells are governed by high-velocity fire ball and pin hole jets. Overall, this study provides critical insights into the jetting behavior and its characterization, which can help us to tune the droplet gasification and the gas phase mixing to achieve an effective combustion control strategy for gel fuels.

show abstract