Influence of Edge-Limited Hot Surfaces on Accidental Ignition and Combustion in Ship Engine Rooms: A Case Study of Marine Diesel Leakage

Abstract

To extend initial ignition-related fire prevention in ship engine room, this work presents a case study of marine diesel leakage for identifying accidental ignition by hot surface. Based on a self-designed experimental platform, a full-scale innovative experimental arrangement was conducted for diesel leakage-related hot surface ignition (HSI) tests in a ship engine room. A series of parameters (e.g., heat transfer, evaporation mode, ignition position, ignition delay time, flame instability, and combustion behavior) for improving the initial HSI of diesel leakage on an edge-limited hot surface were analyzed. A transient sequence corresponding to a change in leakage flow rates ranging from 7.5 mL to 25 mL was tested, and hot surface temperatures (HSTs) were adjusted between 390 °C to 525 °C. Puffing motion accelerated the mixing of HSI-driven vapors with fresh air, which was affected by the edge-based limitation and HSTs. The case study identified the effects of hot surface shape and the most important combinations of HSI-driven combustion characteristics for estimating initial ignition responses. Based on this current work, prediction models were proposed for determining the HSI height of marine diesel for varying leakage flow rates and HSTs. The results indicate that HSI height increases with leakage flow rate and HSI position is influenced by edged hot surfaces, leading the vertical centerline to shift towards the side of the edge structure. The results also revealed that the ignition delay time of diesel leaked onto an edged hot surface decreases as leakage flow rate increases. This change causes the initial HSI to occur earlier, potentially creating an extra risk in ship engine rooms.

1. Introduction

Leaking marine fuel that comes into contact with hot surfaces caused by nearby machinery operating under high load can trigger hot surface ignition (HSI) [1,2,3]. This phenomenon is particularly prevalent in machinery with hot surfaces that have edge structures such as ridges, with adjacent machinery often being the initial source of the leakage. When the machinery surface experiences an initial HSI, the edge structure significantly affects heat transfer, flame spread, and development, which may present significant safety hazards [4]. The fundamental processes induced by the small fuel release in reactive mixtures leads to ignitions with various chemical reaction modes. Real-life HSI scenarios in ship engine rooms involve varying temperatures and reaction conditions. For some edge-limited geometries, it is possible for the hot surface to accelerate the flame, producing a flame velocity-pressure dependence. Possible initial regimes of the chemical reactions ignited by the initial temperature gradient were analyzed by Zel’dovich et al. using a one-step Arrhenius model [5,6]. The one-step model of chemical reaction, which is widely used for studying the initial ignition by a hot source that drives the flame dynamics of obstacle-laden structures, has been used to explain Zel’dovich’s gradient mechanism of flame acceleration [7]. Ongoing work attempts to conduct experimental research to explore how the initial HSI behavior of leaking marine fuel changes with the presence of edge structures.

Previous research has shown significant improvements in the accuracy, sensitivity, and response time of ship fire and smoke detection using multi-criteria approaches [8]. Su et al. [9] indicated a significant temperature gradient within the ship engine room, even at a distance from the fire source. Wu et al. [10] proposed a probabilistic model based on Bayesian statistics and modelling, which can be used to predict fatalities in a ship fire accident. During a ship fire, flames can be influenced by equipment in cabin rooms to flow around and potentially spread [11]. Liu et al. [12] proposed a new flame propagation phenomenon, which was defined as “continuous leakage-dispersion-flow” and “broken-retract,” whereby mass burning rate increases with enhanced ventilation, and flame length and deviation angle changes under different ventilation speeds [13]. Wu et al.’s work revealed that the ultimate strength of ships remained relatively stable until the temperature reached 200 °C [14]. Ship engine rooms commonly utilize diesel engines due to their high dependability and efficiency for heat transmission. In fact, shipping engineering equipment accounts for over four-fifths of the overall transportation in today’s industry [15]. It is noteworthy that the entire combustion process for a pool fire of marine fuel consists of initial growth, steady burning, and decay [16]. The burn rate of marine diesel was about 0.044 kg/(m²·s), which is higher than the experimental burn rate observed in confined spaces [17]. Experiments have demonstrated that the combustion rate and growth induced by a marine fuel can be accelerated in some cases by the thermal insulation of confined compartment [18,19]. Det Norske Veritas [20,21] showed that 63% of fire accidents and explosion disasters on ships originate in the engine room, with over 56% of incidents involving the main engine equipment and pipeline system [22]. However, limited marine fuel leakage and related HSI process in the engine room have received less attention in shipping operations. The combustion of a marine fuel film lasts nearly 100.0 ms because it burns while evaporating, much longer than the duration of an impinging ignition [23]. 3D numerical models can be used to simulate the ignition of evaporated marine fuel above a heated surface [24,25], revealing various combustion features. Relevant studies found that increased ventilation velocity results in elevated burn rates and temperature decreases at the tops of ship engine rooms while triggering a decrease in temperature [26,27,28]. Additionally, obstacles in fuel pools primarily affect the air entrainment of fire plumes, resulting in flame oscillation behaviors [29,30]. These results suggest that the spacing distance of obstacles in marine fuel pools mainly determines their influence on combustion development [31].

Our literatures review indicates that the initial risks, mechanisms, and combustion behaviors of HSIs that occur when leaking marine fuel contacts edged-limited and hot surfaces are not yet fully understood, and real-life scenarios-based experiments remain limited. I this study, we conducted full-scale experiments involving marine diesel leakage ignition tests in a ship engine room in an effort to determine relevant parameters for improving the initial and accidental ignition of marine diesel on edged and heated surfaces. We investigated thermochemical characteristics, such as heat transfer, evaporation mode, ignition position, ignition delay time, flame spread, flame instability, and risk consequence. We present experimental results for the non-premixed combustion of marine diesel vapor on an edged-limited surface. The leakage flow rates ranged from 7.5 mL to 25 mL, and HSTs were varied between 390 °C to 525 °C. In addition, we measured the occurrence heights of marine diesel initial HSIs caused by elevated HSTs of surfaces with restrictive structures, which we then applied to the development of a dynamic predicted model. The initial HSI characteristics and combustion values were compared with those obtained from our experimental tests, resulting in reasonably good agreement, as discussed below.

2. Materials and Method

2.1. Experimental Marine Diesel and Arrangement

Research on shipping enterprises has found that the safe operation of diesel main engines in ship engine rooms remains a primary concern. On-site monitoring reveals that diesel main engine and system piping leaks typically result in leakage rates of 7 to 20 mL. The goals of our current experiments are to elucidate the process of marine diesel HSI under different leakage scenarios and to obtain several index parameters for characterizing HSIs.

Table 1. Physicochemical parameters of the experimental marine diesel used in this case study [32,33].

Material	Density (15 °C, kg/m3)	Kinematic Viscosity (40 °C, mm2/s)	Flash Point (°C)	Pour Point (°C)	Surface Tension (mN/m)	Heat Value (×107 J/kg)
Marine diesel	843	3.5	65.2	23.1	26.9	3.29

presents the properties of the marine diesel used in experiments.

Figure 1 shows a stationary experimental vessel primarily used for identifying ship fires and monitoring ship engine room equipment. The exterior of the ship’s nacelle laboratory measures 6.8 × 6.5 m. The ventilation ducts are located at the top of the laboratory, which stands approximately 3.5 m above the ground. The anemometer (TESTO 440), which is equipped with a 16 mm impeller, has a range between 0.6 m/s and 50 m/s,. It has a ventilation velocity accuracy of 0.01 m/s with a measurement error of ±0.2 m/s in the engine room. Upon entering the ship engine room, one can observe the diesel main engine, oil distributor, seawater pump, air compressor, and other main equipment. The humidity controller (HQ-JS130H) is installed in the laboratory to adjust the engine room’s humidity between 70% and 100% RH (humidity accuracy ±3% FS). The ship sloshing simulator can provide six degrees of freedom of tilt angle to reproduce sloshing behavior under different wave conditions. The frequency of the simulator is set to 0.2 Hz, and the slosh amplitude is less than 5° since the simulated ship in which the laboratory is located is in a fixed mode. Additionally, a multi-channel data acquisition instrument (TOPRIE TP700) is used in the data acquisition module.

This experiment employs a hot surface simulation test rig independently designed and developed by our research team (see in Figure 2). The maximum temperature of the hot surface device can be adjusted up to 750 °C. The hot surface simulator has a measurement accuracy with an error of ±1 °C. The type of edge structures is the same as that frequently found on the surfaces of the main equipment in a ship engine room. The experiments are conducted with a high-temperature device measuring 0.4 m × 0.4 m. The peristaltic pump nozzle is positioned at the top center of the hot surface, and the controlled leakage of marine fuel spills and contacts the heated surface simulator directly. Ten thermocouples (DIN EN 60584) are arranged vertically upwards at 5 mm intervals, with the closest ones being mounted at a height of 5 mm above the hot surface. The data collected is saved in real-time by the collector during the experimental process.

2.2. Heat Transfer and Ignition Mechanism of Leaking Marine Fuel on Hot Surface

When a pipe is damaged, liquid fuel may leak in the form of droplets onto the device’s surface. The droplets then evaporate, forming a vapor mixture around them that is ignited by the hot surface. Gas–liquid equilibrium is assumed to be maintained at the surface of the fuel droplet, with the vapor maintaining a saturated pressure with temperature [34,35]. Based on this assumption, total heat transmission is consumed in the fuel evaporation process, and the relevant energy equation can be presented as Equation (1).

$$2\mathrm{\pi }{r}^2\left(T_s-T_f\right)h=\lambda E+m_f{C}_{p,l}\frac{d{T}_f}{dt}$$

(1)

where T_s is temperature of the hot surface (K); T_f is temperature of the liquid fuel (K); r is radius of the liquid droplet (m); h is the coefficient of heat transfer (W/(m²·K)); λ is the latent heat of the liquid fuel (kJ/kg); E is evaporation rate (kg/(m²·s)); m_f is evaporation mass of the liquid fuel (kg/(m²·s)); C_p,l is specific heat of the liquid fuel (kJ/(kg·K)); and t is the time scale (s).

Assuming isothermal flow of the liquid droplet on the hot surface at low Reynolds number, with negligible variation of velocity with the time compared with spatial coordinates, the heat transfer radiation to droplet is developed by Gülder et al. In this study, it should be noted that the soot and CO₂ produced during the high-temperature pyrolysis or combustion of marine diesel are considered as major factors in Equation (2) [36].

$$\frac{d{Q}_{rad}}{dr}=\frac{2\mathrm{\pi }r{R}_{rad}\left(T_s^4-T_f^4\right)F_b\sigma {\epsilon }_m}{\sqrt{r^2-R_{rad}^2}}$$

(2)

where σ is the Stefan-Boltzmann constant; Q_rad is radiation from heat transfer to the lower half of the liquid surface (W/(m·K)); R_rad is the radial direction(m); F_b is a configuration factor; and ε_m is the total emissivity of the soot/CO₂/H₂O mixtures with constant absorption coefficient.

It is noted that the total emissivity of diesel soot/CO₂/H₂O mixtures in Equation (2) can be further expressed by a three-gray gas model [37,38] with the constant absorption coefficients shown below.

$${\mathsf{\epsilon }}_m={\sum }_{n=1}^3{a}_{g,n}\left[1-\exp \left(-k_{g,n}pL\right)\right]$$

(3)

where a_g,n is the gray gas weighting coefficient; k_g,n is the absorption coefficient (Pa⁻¹·m⁻¹); p is partial pressure (Pa); and L is path length (m).

The radiative flux profile above the leaking fuel surface is essential for estimating the total radiation heat feedback. It is noted that the burn rate of combustible fuels increases with reduced heat of gasification. The heat balance equation of the gaseous diesel phase includes an Arrhenius term in order tos describe the sharp increase in the flame temperature. Equation (4) is developed to correlate the burn rate based on the heat of gasification and smoke point of combustion fuels [39].

$${\dot{ q}}_{fs}^{″}={\dot{m}}_b^{″}\Delta H_g={\dot{ q}}_{c0}^{″}+{\dot{q}}_{r0}^{″}{Y}_s^{0.25}\left\{1-\exp \left[-{\left(\frac{\Delta H_{g}S}{\Delta H_{ch}}\right)}^m{\left(\frac{D}{D_0}\right)}^n\right]\right\}$$

(4)

where q_fs is the heat flux received by the fuel surface (kW/m²); m_b is the area-specific mass burn rate (kg/(m²·s)); ∆H_g is the heat of gasification of the combustible fuel (kJ/kg); q_c0 is the heat feedback by convection (kW); q_r0 is the heat feedback by radiation (kW); Y_s is smoke yield (mass of smoke/mass of fuel) with the factor 0.01 accounting for the presence of radiating gases for flames having very little soot; ∆H_ch is the chemical heat of combustion (kJ/kg); S is the stoichiometric air-to-fuel mass ratio; m, n and D₀ are adjustable parameters in the empirical correlation; and D is the diameter of the pool’s side length (m).

Figure 3 illustrates the ignition process of marine diesel on an unrestricted high-temperature flat plate. The experiment uses marine diesel as the leakage, with a volume of 15 mL. Figure 3 shows that the marine diesel undergoes a phase change due to heat transfer upon contacting the hot surface, forming a flammable vapor above it. As the gas/air mixture increases, the necessary conditions for ignition gradually become available. HSI behavior is influenced by an engine room’s environmental factors, such as ventilation airflow organization, ambient humidity, and thermal feedback from bulkhead surfaces. The presence of an edge structure can affect the initial ignition behaviors of marine diesel leakage.

Heat transfer from the surface varies significantly between these modes; similar occurrences are observed both in Figure 3 at 2199.5 ms and in Figure 4 at the 1544.5 ms. During the natural convection phase, the system experiences convective heat transfers in a single-phase flow state, with a few bubble cores forming on the thermal source. This is attributed to hot surface roughness, which contains numerous small cavities that trap gases. The initial bubble core is formed by a small amount of air. During this stage, heat flow density and heat transfer coefficient are nearly ten times smaller than those during two-phase thermal transmission. Bubble nuclei are produced at temperatures slightly higher than the boiling point. As the HST rises, the fuel absorbs a greater quantity of heat, leading to an increase in the number of bubble cores generated in the area between the fuel and the surface. Additionally, the generation and movement of bubbles within the liquid cause an increase in turbulence. The heat flow density and heat transfer coefficient experience a significant increase, leading to enhanced heat transfer performance and reaching of the core boiling stage. In the case of marine diesel leakage, the nuclear boiling state occurs at HST. Bennett et al. [40] developed the boiling heat transfer coefficient, as shown in Equation (5).

$$h_n={ h}_{n,max}{\left[\frac{T_{s }{- T}_{sat}}{{\left(T_s {- T}_{sat}\right)}_{max}}\right]}^2$$

(5)

where h_n is the heat transfer coefficient for the nuclear boiling mode (W/m²∙K) and T_sat is the saturation temperature of the marine fuel (K).

Assuming (T_w − T_sat)_min = 0.3 (T_w − T_sat)_max, Equation (5) is used to describe the peak heat flux in the nuclear boiling stage [41].

$$q·A_{n,max}^{-1}=143·H_{fg}{\rho }_L^{0.6}{\rho }_v^{0.4}$$

(6)

where ρ_L is fuel density (kg/m³); q is heat flux, W/m²; ρ_v is vapor density (kg/m³); H_fg is the latent heat of vaporization (J/(kg∙K)); and A is surface cross-sectional area (m).

When the HST is high, vapor bubbles are produced more frequently from the marine diesel liquid than from the surface. As a result, some of the vapor bubbles connect to form a small film of vapor but not yet a large film that covers the entire hot surface. It involves the alternating boiling of bubble nuclei and films in time and space. Transition boiling mode begins at a specific surface temperature above saturation temperature, and the heat transfer process can be expressed by Equation (7) [42].

$${\left(T_s-T_{sat}\right)}_{min}=0.127\frac{H_{fg}{\rho }_v}{k_{vf}}{\left(\frac{\sigma }{{\rho }_L}\right)}^{\frac{1}{2}}{\left(\frac{g{\mu }_{vf}}{{\rho }_L}\right)}^{\frac{1}{3}}$$

(7)

where μ_vf is viscosity of the vapor in the film between the surface and the fuel (Pa∙s); σ is fuel surface tension (mN/m); and k_vf is heat conductivity (W/(m·K)).

As the interface between diesel leakage and hot surface enlarges, the liquid wets the surface, and the heat conduction rate achieves its peak value. If heat flow transfers from surface to a critical point, the numerous bubble columns converge and fuse near the hot surface. This leads to the formation of a thin layer of vapor, indicating the onset of the film boiling stage. Figure 3 illustrates the film boiling phenomenon, with the change in diesel leakage on the hot surface at 2674.5 ms. h_F at this stage is calculated using Equation (8) [43] for the heat transfer coefficient.

$$h_{F }=0.425 {\left[\frac{{k_{vf}}^3{H}_{fg}{\rho }_v{\rho }_L^{\frac{3}{2}}g}{\left(T_s-T_{sat}\right){\mu }_{vf}{\sigma }^{\frac{1}{2}}}\right]}^{\frac{1}{4}}$$

(8)

where h_F is the heat transfer coefficient of film boiling mode (W/m²∙K).

3. Results and Discussion

3.1. Effect of Edge Structure on Diesel HSI Position with Varying Leakage Rate

Figure 5 displays the HSI height of diesel leakage on an edged and hot surface for varying amounts of 7.5 mL, 10 mL, and 15 mL. HSI height above the hot surface increases with the amount of diesel leakage. The liquid medium undergoes a rapid phase change during the initial stage of marine diesel leakage on the hot surface. When the HST reaches 450 °C, the marine diesel typically undergoes film boiling, forming a thin gas-phase interface between the hot surface and the liquid fuel. This causes a change in the heat transfer mode. Figure 5a illustrates the HSI process of a hot surface when diesel leakage is minimal, with approximately 7.5 mL of diesel leakage. In the case of a small leakage of marine diesel, a large amount of evaporation products is not generated when the diesel contacts the hot surface. In scenarios with less leakage, the specific surface area of diesel leakage on the hot surface is relatively large. A high-temperature surface with a constant output of high-temperature continuously heats the leaking material, resulting in the formation of a certain amount of white combustible gas/air mixture in the vertical space at around 4250.0 ms. The ignition point occurs 5988.7 ms after the marine diesel contacts the hot surface. Shortly thereafter, at 6005.3 ms, a fire nucleus is formed by igniting the leakage near the hot surface. Within 8.4 ms, the initial fire nucleus grows to over five times its original size.

As the amount of marine diesel leakage increases to 10 mL, the entire initial ignition process can be observed in Figure 5b. Due to the increase in diesel leakage, the quantity of evaporation products generated in the vertical space at the same time as leakage occurrence is significantly higher than in the experimental scenario with less diesel leakage. At 2250 ms after the marine diesel contacts the hot surface, white products and mixtures have already been generated in a certain area. Influenced by the movement of the airflow organization, the gas phase medium produced by the heat spreads rapidly to the top of engine room. At 2450.5 ms, initial ignition behavior at the hot surface of the fuel occurs. Subsequently, the initial fire nucleus gradually become clear after 24.9 ms. After about 8.3 ms, the flammable gas mixture around the fire nucleus begins to ignite, and a bright orange-red light is emitted. Figure 5c shows that at 1083.3 ms after the leaking diesel (15 mL) contacts the hot surface, more flammable medium is generated in the vertical space. Subsequently, white flammable evaporation products are continuously produced from the hot surface and diffused into the air, which are then mixed with air. As a result of significant diesel leakage, vapor/air mixtures spread extensively above the hot surface, not only in the vertical space but also horizontally. At 1905.6 ms, initial ignition occurs, followed by the rapid expansion of the initial fire nucleus in all directions. The flame spreads rapidly and emits a bright yellow glow due to the surrounding flammable mixture. The comparison demonstrates that the ignition position of diesel leakage is influenced by the presence of an edge structure on the hot surface, causing the vertical centerline to shift towards the side where an edge-limited structure is present.

The evaporation products of the diesel on an edge-limited surface mix with fresh air in the engine room to form a stable combustible vapor/air mixture in the upward field. This leads to the vertical flow of higher-temperature combustible gases, while the surrounding air fills in below to create upward air convection. The plume dynamics depend on air flow and air velocity in the ship engine room and the plume buoyancy flux [44]. Obviously, it is important to understand the conditions under which the combustible vapor/air mixture from marine diesel leakage influences plume dynamics and the near field of edge-limited hot surfaces. As a result of continuous heat transfer from the hot surface to the marine diesel’s surface coupled with thermal feedback around the ship’s nacelle, the marine diesel exhibits initial HSI behavior after a certain time point. Figure 6 illustrates the corresponding evolution of the HSI height of diesel leakage with the changed flow rate in engine room. Overall, the height of the ignition location on the hot surface in the vertical space increases as the amount of marine fuel leakage increases. This indicates that in the event of a small aperture leakage, the initial ignition location will occur above the equipment surface as the HST rises to the HSI temperature of the diesel. The initial monitoring and prevention of fires in engine rooms can be hindered by unfavorable conditions. This study proposes a model for improving the prediction of HSI height for different leakage flow rates based on experimental data. Equation (9) shows the fitted relationship of this model, which has R² value of 0.982.

$$H=H_{0 }+ \frac{H_{ave}}{H_{max}}\ln {\left(\frac{T_s-T_{sat}}{T_0}\right)}^2-0.0151$$

(9)

where H_max is maximum HSI height (m); T_s is the HST in the ship engine room (K); T₀ is the initial HST of the equipment (K); H_ave is mean HSI height of diesel leakage above an edge-limited surface (m).

According to Equation (8), it can be further developed as following expression.

$$H=H_{0 }-\frac{6.85H_{ave}}{H_{max}}\ln {\left[\frac{{k_{vf}}^3{H}_{fg}{\rho }_v{\rho }_L^{\frac{3}{2}}g}{h_F^4{\mu }_{vf}{\sigma }^{\frac{1}{2}}{T}_0}\right]}^2-0.0151$$

(10)

3.2. Edge-Limitied HSI Position of Marine Fuel for Elevated HSTs

The experiment assumes a 15 mL diesel leakage in an engine room, with increasing HSTs from 390 °C to 525 °C. Tests were repeated to obtain valid data. The ignition position exhibits a pattern of rapid decrease followed by gradual stabilization as HST increases. Figure 7a indicates that although HSI can occur, the ignition process requires a longer gestation period due to the relatively low temperature of the hot surface. When diesel leaks and contacts a hot surface containing edge structures, it flows around the surface, and some of it aggregates at the junctions of the edge structures. Due to continuous heat transfer and convection, the marine diesel undergoes a phase change and produces combustible gas-phase media. Initial ignition occurs in the air at 4792.8 ms, and the HSI of diesel leakage appears at a higher vertical height when exposed to a temperature of 390 °C. This location is not only farther away in the vertical space but also indicates greater horizontal offset. The addition of an edge structure to the hot surface significantly affects the ignition position, which is jointly influenced by the airflow organization in the ship’s nacelle and the thermal feedback of the bulkhead. When the HST gradually increases to 410 °C, the Leidenfrost effect appears after the diesel leaks and contacts the hot surface due to the increase in HST. The marine diesel on the hot surface exhibits rapid diffusion, resulting in the generation of a gas-phase interface. As shown in Figure 7b, there are few gas-phase products in the air at 2190.9 ms. As time progresses and the diesel continues to contact the hot surface, the amount of combustible mixture in the vertical space increases. At 3220.1 ms, the ignition position decreases as HST increases from 390 °C. Figure 7c illustrates the ignition process for 15 mL of the diesel leakage onto an edge-limited surface when the HST increases to 430 °C. The HSI-driven ignition position drops compared with that for a HST of 390 °C. It quickly produces a large amount of white smoke that spreads vertically through the ventilation airflow. Additionally, the evaporation products of diesel spread horizontally. The ignition of diesel leakage occurs at 2431.7 ms, forming an initial fire nucleus above the structural junction near the hot surface’s edge structures. As the HST temperature increases to 450 °C, the HSI behavior of the diesel leakage exhibits differential changes, as shown in Figure 7d. At 1086.2 ms, a significant amount of diesel vapor accumulates above the hot surface. It is noteworthy that the presence of an edge structure results in more diesel vapor appearing above it. U-shaped gas mixture diffusion phenomenon is visible above the hot surface. As contact time increases, the vertical space is filled with a flammable vapor mixture. At 1913.9 ms, initial ignition occurs near the hot surface. The fire nucleus due to the initial ignition then increases instantaneously and exhibits a distinct yellow light. As the HST increases further, the overall ignition height levels off and no longer decreases significantly. However, there is still a shift in the location where the HSI occurs.

Experimental data reveal that the ignition height of the diesel leakage decreases as the HST increases, regardless of the presence of an edge-limited structure. The data show significant differences in ignition position deviation due to the influence of the edge structure. When the HST with no edged surface is 480 °C, the ignition height of diesel leakage is 0.465 m. However, if there is an edge structure on the hot surface, the ignition height of diesel leakage reduces to 0.109 m. It is observed that the presence of an edge structure affects the acquisition of the screen, resulting in a larger vertical ignition position of the leaking marine diesel. Simultaneously, it is seen that the ignition point of the marine diesel shifts vertically when an edge structure is present. The ignition point is located in close proximity to the interface of the hot surface edge. The shift in HSI position is caused by the thermal feedback effect generated by an edge-limited mechanism, which results in the flammable gas mixture above the region reaching ignition conditions. It is also noteworthy that ignition of the marine diesel occurs at a relatively low temperature on hot surfaces where edge structures are present. The thermal conductivity of the hot surface is enhanced by the presence of edge structures that facilitate increased heat transfer within the edge junction region. This, in turn, results in marine diesel HSI occurring closer to the location of the edge structures. Figure 8 illustrates the relationship between the HSI height of leaking diesel and HST in the presence of an edge structure. It is observed that at lower HSTs, the marine diesel requires more time to generate combustible vapor above the hot surface. The diesel leakage has an initial ignition height of 0.407 m when exposed to a HST of 390 °C in the presence of an edge structure. As HST increases, the phase change of diesel leakage accelerates, producing the required gas mixture for ignition in a shorter time. When HST reaches 410 °C, the ignition height above the surface exceeds 0.179 m, although it still decreases significantly compared to the ignition height at 390 °C. Several experimental tests have verified that the HSI height of diesel leakage falls within the range of 0.13 m to 0.15 m as HST increases between 430 °C and 450 °C. As HST exceeds 470 °C, the HSI height of diesel leakage remains stable between 0.10 m and 0.12 m. The ignition location of diesel leakage remains consistently 0.13 m to 0.15 m in the vertical space above the surface. These results were verified in multiple experimental tests. The experimental data enable determination of the vertical height of the ignition region, leaving the possibility of the ignition position of diesel leakage shifting horizontally and producing varying results for each test. Equation (11) is derived in this case study, providing the correction between HSI height and HST with an R² value of 0.968. It is important for predicting the occurrence of HSI of marine diesel in case of leakage, and flow based on monitored HST is crucial for preventing initial ignition accidents.

$$H=H_0+\frac{H_{ave}}{3}\exp \left[-\frac{0.078{k_{vf}}^3{H}_{fg}{\rho }_v{\rho }_L^{\frac{3}{2}}g}{{h_F^4{\mu }_{vf}{\sigma }^{\frac{1}{2}}T}_0}\right]$$

(11)

where H_ave is the mean HSI height of diesel above an edge-limited surface (m).

3.3. Influence of Edge Structure on Ignition Delay Time of Marine Diesel Leakage

Previous studies [45,46,47] indicate that ignition time is generally short when HSI occurs. However, it is worth noting that there is a lack of significant discussion and in-depth knowledge from the HSI regarding the ignition delay time. This term refers to the time it takes for the fuel vapors above a hot surface to reach ignition conditions and then absorb enough heat from the environment to increase their temperature to the HSI temperature. It serves as an important indicator for characterizing the leakage of marine diesel onto edged hot surfaces and its ignition behaviors. Experimental results using marine diesel are shown in Figure 9 along with a comparison to previous experimental data. HSI behaviors of marine diesel on hot surfaces and the required ignition gestation period show a general decreasing trend with an increase in the HST. For localized ignition of marine diesel on a hot surface, two basic prerequisites are necessary: a suitable fuel vapor/air mixing ratio in the ignition generating area and the formation of an initial flame. However, this study emphasizes the significance of environmental factors within the engine room, particularly the nature of equipment surfaces that generate high temperatures. The ignition delay time of diesel leakage varies to a greater extent when the hot surface of the equipment is a high temperature flat plate and depends on HST. When the HST is relatively low, the probability of ignition of leaking diesel on a flat plate with no edge structure is low as well. As the HST increases to 475 °C, the ignition probability reaches 50%, and the corresponding ignition delay time is approximately 3.692 s. When the HST increases to 485 °C, the ignition delay time shortens to 1.425 s accordingly. The ignition delay time of marine diesel on an unconstrained hot surface stabilizes between 500 °C and 525 °C. When edge structures are present on the equipment surface, the time required for the marine diesel to flow and ignite above them changes. In the presence of the edge structures, marine diesel has a greater than 50% probability of ignition at relatively low HSTs. Figure 9 illustrates that the ignition delay time of marine diesel is approximately 4.792 s when the HST of the edge structures reaches 390 °C. As the HST increases, the time required for initial ignition of marine diesel above the hot surface gradually decreases. At a HST of 420 °C for an edged hot surface, the ignition delay time of diesel leakage in an engine room goes down to 2.821 s. As the HST increases, the time required for the ignition of diesel leakage in the presence of an edge-limited structure gradually stabilizes. When the HST exceeds 450 °C in an engine room, the ignition delay time remains steady between 1.78 s and 1.91 s.

Corresponding experimental tests were conducted in this study using marine diesel with leakage flow rates ranging from 7.5 mL to 25 mL to obtain ignition delay time data for an edged hot surface. Figure 10 shows that the ignition delay time of the diesel leaking on an edged hot surface decreases as the leakage flow rate increases. This suggests that the more marine diesel leaks on an edged hot surface, the faster the initial HSI occurs. It is therefore crucial to elucidate the rules governing the change in ignition delay time with increasing marine diesel leakage to effectively prevent and control initial fire risk accidents in engine rooms. Small leakages in the system pipeline results in a low volume of leakage due to the low pressure within oil pipelines of a ship’s engine room. For the HSI test in this experiment, a marine diesel leakage flow rate of 7.5 mL was used. As the experimental results indicate, it takes approximately 6.005 s for marine diesel to initially ignite on an edge-limited surface. This shows that the marine diesel does not undergo significant phase change upon contact with the hot surface but instead flows on the surface. The combustible gas mixture formed in the vertical space at a leakage flow rate of 7.5 mL is relatively small, and the diffusion range is narrow. When the diesel leakage increases to 10 mL, HSI delay time shortens to 2.565 s. This reveals that the change in leakage level significantly affects initial HSI behaviors of a flammable liquid. The ignition delay time of diesel leakage decreases to 1.914 s when the flow rate increases to 15 mL. Additionally, the ignition delay time of the marine diesel decreases with increasing leakage at the same HST. When the amount of marine diesel fuel leakage exceeds 17.5 mL, the ignition delay time on the hot surface is less than 1.541 s. As the flow rate of the leaking marine fuel increases, the time needed for the initial HSI phenomenon to occur on an edged-limited surface gradually becomes smoother. The flow rate of diesel increases with the larger leakage caliber of the main engine pipeline due to the high viscosity and poor fluidity of marine diesel. The leakage area of marine diesel onto hot surfaces also increases, which enhances the rate of heat transfer by increasing the surface area for heat transfer. The marine diesel reaches its equilibrium temperature more quickly after the diesel’s evaporation rate with increased contact with the hot surface. As a result, more leakage of marine diesel has less time to accumulate vapor concentration, reducing the risk of HSI in the presence of a marginal structure. This change causes the initial HSI of the diesel vapor to occur earlier, potentially creating a new fire hazard in the ship engine room. In the event of a significant leak in a ship engine room piping system, the time required for ignition of the diesel leakage on an edged hot surface is reduced. Therefore, it is crucial to enhance fire control planning in the area where the HSI is the most likely to occur.

3.4. Flame Spread of Leaking Marine Fuel above Hot Surface Edge Structure

Figure 11 presents the flame propagation transfer for varying diesel leakage flow rates after the appearance of HSI on a bounded hot surface. The HST, which includes an edge structure, reaches approximately 450 °C. The HSI-driven ignition of diesel leakage occurs near the top of the edge interface at this temperature. The flame rapidly expands in all directions in air after the initial ignition occurs. Figure 11a illustrates a small pipeline leakage incident when marine diesel leaks in an engine room with a flow rate of 7.5 mL. It can be observed that the fire nucleus formed in the air changes from an initial blue color to orange-red at 6024.9 ms. Incomplete combustion during the combustion process produces soot. The soot particles emit long-wavelength electromagnetic radiation at high temperatures, causing the entire thermochemical reaction region to appear red or yellow. At 6199.2 ms, the flame formed in the air gradually moves towards the hot surface. At 6232.4 ms, the flame contacts the edge structure of the hot surface and starts to burn steadily. Due to the small amount of leakage, there is a significant quantity of flammable gas mixture formed by heat transfer from the hot surface. Therefore, after the initial ignition of diesel leakage, the flame is transferred to the hot surface within a short time. Ultimately, sustained combustion occurs at the interface of the edge structure.

As shown in Figure 10b, an increase in the leakage flow rate leads to a change in the flame transfer process of leaking diesel on the hot surface. When the leakage flow rate reaches 10 mL, the diesel leaks and meets the high-temperature edged surface, forming an initial fire nucleus. Then, the flame propagates in all directions due to the high concentration of the flammable gas mixture around the fire nucleus. At 2606.5 ms, the flame spreads to the surface, triggering a pool fire phenomenon utilizing the remaining marine diesel. If the combustion occurs on a flat plate, the marine diesel combustion process should be perceived as a stable conical structure. The flame pattern above the conical structure resulting from pool fire combustion of marine diesel is affected by the engine room’s environmental conditions. When the high-temperature equipment surface has an edge structure, the flame structure changes from a cone to a trapezoidal shape due to the higher heat transfer intensity at the interface of the edge structure. At 2623.1 ms, the flame height at the edge structure interface is higher, resulting in a more trapezoidal flame morphology. The marine diesel evaporates more at the center of the hot surface after contact, while some of it flows to the edge of the hot surface. At 3063.3 ms, the flame is induced by the pool fire and forms on the edged surface, which appears as an interrupted surface on the centerline. This is due to the high consumption of combustible vapor/air mixture in the vertical space above the edged hot surface and the insufficient supply of fresh air at the center. At 5164.3 ms, the flame persists at the edge of the hot surface without combustion, and there is no re-ignition at the center of the edge-limited surface.

Figure 11c illustrates the flame transfer process following initial ignition of marine diesel on a hot surface for a leakage flow rate of 15 mL. The bright flame expands significantly by 1930.5 ms. At 1947.1 ms, the flame spreads horizontally to more than 0.3 m after the initial HSI at 1996.9 ms, causing the combustible gas mixture above it to boom. This suggests that a significant amount of diesel vapor/air mixture is generated due to the large amount of marine diesel leakage during the pre-evaporation process, which accumulates in large quantities near the hot surface area. At approximately 2337.2 ms, the flame reaches its maximum size with a height of over 1.15 m. By 2710.7 ms, the combustion process of the diesel leakage on the hot surface shows a clear layering. The structure of the flame consists of a stable combustion zone closest to the hot surface, a flame interruption zone located in the vertical space, and the uppermost buoyant plume zone. The burning flame continuously releases heat to maintain density differences within the flame, while the density differences within the non-burning plume decrease as it mixes with air. The plume that does not burn creates a significant difference in density in the vertical space above the edge-limited surface, causing the flame to zigzag. In contrast, a flame that burns on a hot surface will have a meandering pattern due to the relatively small density difference. This pulsating flame phenomenon is more likely to occur near edge structures. As the combustible gas mixture above an edged hot surface is consumed, the flame height gradually drops, resulting in sustained combustion on the hot surface. At 5710.7 ms, the flame develops steadily at the edged hot surface and shows an intermittent flame phenomenon at the center of the edge-limited hot surface.

To clarify the effect of edge-structured hot surfaces on flame development in leaking marine diesel combustion, this case study further analyzed experimental data for different HSTs. The flame is turbulent under all conditions where the temperature of engine room equipment changes. The turbulence pattern of the flame characterizes the instability of the rising plume gas, which exhibits a density gradient, and the cold air with suction under gravity. Above an edged hot surface, the flame becomes unstable due to fresh air being drawn into the flame edge along the edge structure. The ignition delay time before initial ignition is long due to the relatively low temperature of the hot surface, as shown in Figure 12a. During this time, a flammable gas mixture collects in the vertical space above thermal source. The flame expands in all directions while traveling along the path between the initial fire core and the hot surface, and it develops entirely within the vertical area above the thermal source between 4851.1 ms and 5747.5 ms. When the leakage flow rate of diesel is 15 mL, the remaining diesel on the hot surface immediately ignites when it meets the downward flame, resulting in pool fire combustion. At 6328.5 ms, the conical pool fire on the surface disappears due to the edge structure. The unstable plume gradually descends to a position near the mouth of the oil pool, and starts to tumble and expand upward. The plume below it stretches and gradually breaks away from its main body. This phenomenon occurs due to the puffing motion, which accelerates the mixing of HSI-driven products and vapors with the surrounding fresh air. Constant swirling of fuel vapors from the plume to the surrounding air and the swirling of the surrounding air into the plume facilitate this mixing.

When the HST reaches 410 °C, the ignition height appears at a location closer to the edge structure, as shown in Figure 11b. At 3253.3 ms, the flame formed by the initial fire nucleus rapidly expands in all directions. When the flame meets the hot surface, it ignites the remaining diesel material, resulting in a pool fire. At 3328.9 ms, the marine diesel triggers a booming ignition on the hot surface, causing the flame to spread to its maximum range. The combustible gas-phase medium is further consumed as diesel vapor/air mixture in the initial HSI region and continues to react chemically. At 3818.6 ms, a change in the flame propagation range is observed. This change is highlighted by a gradual decrease in the flame extent. Such decrease is due to the gradual depletion of combustible vapor generated by the phase change under the hot surface action in which the initial HSI occurs. At 4885.3 ms, the flame propagates towards the hot surface in a more stable pattern. As the marine diesel above the hot surface undergoes a phase change, the flame spreads to areas where the fuel gradually touches the hot surface. This results in a steady combustion of the pool fire for a period of time. During this period, the flame on the hot surface exhibits intermittent behavior in the middle, while continuing to develop at the edge structures.

Figure 12c illustrates the flame transfer process that occurs after ignition of the hot surface of diesel leakage at a temperature of 430 °C. At 2464.9 ms, the flame rapidly expands in all directions and approaches the hot surface. At 2514.7 ms, the vapor/air mixture ignites for the first time, and the flame passes over the hot surface. At 2921.4 ms, a steadily burning pool fire gradually develops on the hot surface. The presence of the fringe structure causes the air to exhibit coil suction behavior on the side of the flame root. This is due to the lateral pressure induced by the lateral velocity component of the air sweeping up at the root of the plume. The intensity of this behavior is less than that of the burning region above the edge structure and is largely confined. The figure shows the density and velocity differences at the edge structure of the hot surface, which are small compared with those of the surrounding area. The Rayleigh-Taylor instability has less effect in this region. However, the heat released from the combustion reaction at the edge structure enhances the density and velocity differences between the plume and the surrounding air, leading to further instability of the plume above the edge-limited surface.

As the HST increases to 450 °C, the ignition and combustion process of the marine diesel changes over time, as shown in Figure 12d. At 2326.9 ms, the initial ignition generates a fire nucleus that has expanded to a larger scale. At 2464.9 ms, the combustion flame is transferred directly to the edged hot surface. The remaining diesel is ignited, resulting in pool fire combustion. By comparing the combustion process of marine diesel on hot surfaces at different temperatures, it is observed that the puffing phenomenon is affected by the edged hot surface and also that it is correlated with the temperature of the surface itself. At 2863.3 ms, obvious flame pulsation is detected. The flame pulsation is a significant finding regarding the fluid movement and interaction within a flame during the ignition process of hot surfaces typically found in ship engine rooms. This finding is crucial for preventing ship fire accidents. The puffing phenomenon becomes less pronounced as the HST of an edge-limited surface increases. The air continuously feeds the interior of pool fire, causing an increasing amount of vapor to burn on the edge of the hot surface. The relevant process can affect and disrupt the puffing formation process.

4. Conclusions

In the current study, relevant experimental tests were conducted, and the parameters for HSI and combustion were analyzed. Noteworthy findings can be summarized as follows:

• The presence of edge structures on hot surface affects the plume swirling process, which in turn exerts influence on the flow state of a plume. The varying combustion patterns and turbulence of plumes on hot surface impact the radiative and convective thermal feedback effects of the flame on marine diesel.
• HSI position is influenced by edged hot surfaces, causing the vertical centerline to shift towards the edge structure. Equation (9) is a prediction model for determining HSI height of marine diesel for varying leakage flow rates. Equation (11) was developed to evaluate the occurrence of diesel HSI in cases of leakage based on monitored HST, which is crucial for preventing initial ignition.
• Ignition delay time of diesel leakage on edged hot surfaces decreases as the leakage flow rate increases. This change causes the initial HSI of gaseous mixtures to occur earlier, potentially creating new fire hazards in ship engine rooms. In the event of significant leakage from a piping system, the time required for diesel leakage HSI on edge-limited surfaces is reduced.
• Unstable plumes gradually descend to a position near the hot surface and begin to tumble and expand upward. This phenomenon occurs due to a puffing motion, which accelerates the mixing of HSI-driven vapors with the engine room’s air. Puffing motion is affected by edged hot surfaces and is also correlated with the HST.