Smoke Flow Research Articles

Full-scale experimental study on fire-induced smoke movement and control in an underground double-island subway station

The research carried out full-scale fire experiments for smoke movement and control in an underground double-island subway station. Four kinds of fire scenarios were designed in the station hall and platform respectively, and series of thermocouples were arranged in the public space. Several vital fire parameters were analyzed, including the vertical temperature distribution, the longitudinal temperature distribution, smoke layer height and smoke front arrival time. Besides, the effectiveness of exhaust system was evaluated for fire emergency. The results showed that the smoke flow characteristics under different fire scenarios exhibited diversely. Under natural ventilation, the maximum temperature excess near fire source was 18 ℃ and 21℃ in the station hall and platform fire scenarios separately, and the smoke temperature attenuated gradually with the longitudinal distance from fire. Besides, the smoke layer exceeded the height of 4.0 m and 2.8 m when fire in the station hall and platform. After activating the exhaust system in the B-end, it was obvious that the smoke layer height increased and the temperature excess decreased. Additionally, the smoke front arrival time was influenced by the fire source locations and ventilation conditions. In the above scenarios, the smoke diffusion velocity was empirically determined as 0.11–0.18 m/s and 0.21–0.35 m/s when a fire in the station hall and platform respectively. The conclusions from the on-site fire experiment could provide data support for the fire accident emergency in an underground double-island subway station.

Prevention of multiple patterns of combined buoyancy- and pressure-driven flow in longitudinally ventilated sloping multi-branch traffic tunnel fires

Longitudinal ventilation is widely employed for smoke control in tunnels due to its relatively low installation cost and compactness. However, smoke control using a longitudinal ventilation system becomes complex when applied in a multi-branch sloping tunnel. In sloping tunnels, the stack effect and forced longitudinal ventilation may combine to produce multiple smoke flow patterns, even under identical or similar boundary conditions. Furthermore, such flow multiplicity behavior is considerably more complicated in a multi-branch sloping tunnel than in a single sloping tunnel. In this paper, an optimal longitudinal ventilation strategy was proposed to prevent smoke flow multiplicity induced by the combined buoyancy- and pressure-driven effects of a downhill sloping, longitudinally ventilated multi-branch tunnel fire. The ventilation network of the multi-branch tunnel was divided into three basic regions: the branches upstream of the intersection node, an on/off-ramp, and the other branches downstream of the intersection node. Based on this division and a hydraulic analysis, the multi-branch tunnel was characterized as a system consisting of several three-branch structures. The critical fan-induced pressure rise required to prevent smoke flow multiplicity was then derived by a potential analysis method according to the fire location in the tunnel. The optimal strategy was achieved by integrating the fan-induced pressure rises in each tunnel branch. The results reveal that the total pressure at the branch intersection node can be varied to regulate the integrated fan-induced pressure rise in each branch. The efficiency of this strategy in preventing flow multiplicity for all typical fire source locations in a three-branch structure was then demonstrated via numerical simulations. The proposed approach has implications for improving tunnel emergency ventilation design and fire protection.

Design and performance test of drum kiln for durian peel carbonization

Durian is one of famous fruit in several places in Indonesia. Only about 20 % of durian fruit is utilized or consumed. The huge amount of waste, especially durian peel, still un-utilized and potential causing some environmental problems. Based on the composition of durian peel, carbonization is one of the best process to utilize durian peel to produce bio-charcoal. The research aims to design a specific kiln for durian peel carbonization and conducting performance test of the kiln. The kiln has been designed as drum type, by using used oil drum as main raw material of the kiln, so it will be easy to find even in the village area and relatively cheaper and easier on kiln fabrication. The performance test shows that the kiln capacity is around 10 kg durian peel. The best performance was achieved in the second test, which produce yield 25% of good quality bio-charcoal in 75 minutes process at temperature range of 326 – 359 °C. The charcoal has LHV (low heating value) of 6347 kCal/kg at water content of 4.66% wet basis, which is compatible to Indonesian National Standard (SNI) quality. Based on the performance investigation, the kiln design should be improved to avoid some weakness performance such as smoke flow and duration time of carbonization.

Flow and spreading behaviors of coaxial jets wake

Flow and spreading behaviors of swirling jets using a dual-blockage disk are studied experimentally. The control and blockage disks are placed concentrically in tandem. The smoke flow patterns are obtained using the flow visualization technique. The axial velocity and turbulence intensity are detected using a 1-D hot-wire sensor. The jet spreading characteristics are illustrated by using an Edge Detection Method. Two pairs of lung-formed vortices and triangle-formed vortices are induced in downstream wake at Rec ≤ 200. Two vortices are found near the field at 200 < Rec < 700, while no toroidal structure was found above the reflected jet at Rec ≥ 700. The recirculation bubble length was increased with increasing Rec until Rec < 700. The axial velocity and turbulence intensity at 200 < Rec < 700 are significantly greater than those in other modes. At Rec ≥ 700, the shear-layer vortices are found far away from the control disk.

Effect of annular flow pulsation on flow and mixing characteristics of double concentric jets at low central jet Reynolds number

This study investigated the effects of annulus pulsation on the flow characteristics and mixing properties of double-concentric jets at a low central jet Reynolds number. The central jet Reynolds number was fixed to Rec = 313. Annulus pulsation was induced by employing a solenoid valve. Streak pictures of instantaneous smoke flow patterns and long-exposure images of jet flows were observed using the laser-light-sheet-assisted smoke flow visualization method. A binary edge-detection technique was used to compute the jet spread widths. Particle image velocimetry (PIV) was utilized to examine the time-averaged velocity vectors, streamline patterns, velocity properties, turbulence intensity distributions, and vorticity contours. A radial dispersion improvement index was defined to quantitatively estimate the mixing property using the tracer-gas concentration detection technique. When the annular flow was pulsated with high amplitudes, it merged toward the central axis owing to an increase in radial momentum. Stagnation points were intermittently witnessed on the central axis. Large jet spread width exhibited, especially at low annulus Reynolds numbers. Turbulence intensities increased drastically in the flow field by the axial elongation of the recirculation region due to the annulus pulsations. The axial elongation of the recirculation region and the intermittent stagnation points led to effective dispersion of the central jet into the annular flow. The radial dispersion improvement index revealed that the mixing capability between the jets could be enhanced up to 80% in the recirculation region by applying annulus pulsations.

Experimental and numerical study of smoke behavior in high-rise stairwells with open and closed windows

This study numerically and experimentally investigates the transport phenomena in buoyancy-driven smoke inside stairwells in a high-rise building. Hot smoke is supplied at the bottom of a small-scale, 2-m high stairwell prototype, and the smoke velocity and temperature are measured and compared with the corresponding numerical results. For all-windows-closed cases, the Fire Dynamics Simulator (FDS) model is used to predict the smoke velocity and temperature fields, which are found to be in good agreement with the experimental data. Obstruction caused by the stairs is observed to slow the smoke flow, which results in staggering and repeated vortical flows in all stairwells, confirmed by flow visualization. The flow path lines and vortex formation of the smoke inside the stairwells are visualized using the laser-induced fluorescence (LIF) method; these vortical structures also corresponded to the results of FDS simulation. Furthermore, the effect of heating power (Q) is investigated in the range of 60–180 W for experiments and 1–4 kW for simulations. Both temperature and velocity increase with Q. Having one open window at various building heights is shown to have small effect on the overall smoke temperature, although having many open windows causes a temperature drop owing to the inflow of fresh, cool air. Having one open window at various building heights slightly slows the smoke velocity, although the velocity is significantly decreased when many windows are open. Therefore, the intake of fresh air slows the overall smoke dynamics. Moreover, the effect of Q in the range of 2–20 MW over building heights of 60, 120, and 240 m is numerically simulated. The rate at which the smoke reaches high elevations is determined for all-windows-closed and all-windows-open cases based on our parametric studies. The smoke rise time (t) is shown to be proportional to ~ Q-1/3 for all building heights, which is the same time scale as the one predicted by the plume theory. However, because of the complex internal geometry of confined buildings including stairwells and corridors, the magnitude of the smoke rise time for the building is much larger than that predicted by the plume theory. Therefore, the current experimental and numerical findings may be useful as design guidelines for building safety engineers.

Mass flow rate through a horizontal opening at small pressure differences

The smoke and air flow through a horizontal opening was investigated using a model-scale room equipped with a horizontal opening on the ceiling. A smoke layer was created under the ceiling using a small gas burner. The pressure difference across the horizontal opening was changed by controlling the air supply to and exhaust from the space above the opening. Smoke and air flow patterns were either downward uni-directional flow of air, bi-directional flow of air and smoke or uni-directional flow of smoke upward. By changing the pressure difference across the opening, critical pressure differences to cause uni-directional flow were determined. It was found that the critical pressure was 0.98 of the smoke layer buoyancy for the onset of upward uni-directional smoke flow, and 0.47 for downward uni-directional air flow. From the measured carbon dioxide concentration, the mass flow rates of smoke and air were calculated and correlated with the non-dimensional pressure difference in the case of bi-directional flows.

A simple model for predicting the smoke spread length during a fire in a shallow urban road tunnel with roof openings under natural ventilation

When a fire occurs in a road tunnel with roof openings, smoke generated by the fire is exhausted through the roof openings in the ceiling by natural buoyancy of the hot smoke, and as a result, the smoke spread length in the tunnel is limited. In this study, we conduct fire experiments using a 1:10 scale model tunnel with roof openings of a 10% opening ratio, and propose a simple model for predicting the smoke spread length in the tunnel based on the results of the experiments. A constant smoke spread length is both observed in the fire experiments and predicted by the simple model. Sensitivity analysis on the effects of the parameters used in the simple model on smoke spread length clarifies the dominant parameters. Furthermore, we clarify that the smoke spread length is decided by a balance between inertia and buoyancy in the hot smoke flow and the inertia of fresh cold air flow induced by the fire.

Study of Effectiveness of CO and Smoke Alarm in Smoldering Fire

Study of Effectiveness of CO and Smoke Alarm in Smoldering Fire

Performance of Different Front-Opening Unified Pod (FOUP) Moisture Removal Techniques With Local Exhaust Ventilation System

A highly effective moisture removal technique for front opening unified pods (FOUPs) can significantly improve the production yield of semiconductor manufacturing industries by minimizing the number of faulty dies rejected due to humidity related defects in silicon wafers. In this study, FOUP purging using clean dry air (CDA) was applied using conventional and diffuser purge method. An innovative laminar air curtain was also used to increase the performance of the FOUP purging techniques. These moisture removal techniques were applied in an open FOUP connected to a mini-environment in which a local exhaust ventilation (LEV) was installed. A novel smoke flow visualization technique and particle image velocimetry (PIV) were employed to visualize the flow pattern inside the FOUP and to generate a velocity vector field from the flow field. In addition, digital humidity sensors were installed inside the FOUP to evaluate the performance of the techniques. The flow analysis and humidity data show that the air curtain can significantly improve the performance of the purge methods. The observations demonstrate that the application of the air curtain, in combination with diffuser purge, can significantly improve the FOUP cleanliness. The LEV system was not found to have influence on the flow field inside the FOUP.

Smart Detection of Fire Source in Tunnel Based on the Numerical Database and Artificial Intelligence

The fire event in a tunnel creates a rapid spread of heat and smoke flows in a long and confined space, which not only endangers human life but also challenges the fire-evacuation and firefighting strategies. A quick and accurate identification for the location and size of the original fire source is of great scientific and practical value in guiding fire rescue and fighting the tunnel fire. Nevertheless, it is a big challenge to acquire fire-source information in an actual tunnel fire event. In this study, the framework of artificial intelligence (AI) and big data is applied to predict the fire source in a numerical model of the tunnel. A big tunnel fire database of numerical simulations, with varying fire locations, fire sizes, and ventilation conditions, is constructed. Temporally varied temperatures measured by multiple sensor devices are used to train a long-short term memory recurrent neural network. Results demonstrate that the location and size of the tunnel fire and the ventilation wind speed can be predicted by the trained model with an accuracy of 90%. Sensitivity analysis is also carried out to optimize the database configuration and spatial–temporal arrangement of sensors in order to achieve a fast and reliable fire prediction. This work addresses the possibility of AI-based detection and prediction of fire source and hazard, thus, providing scientifically based guidance for smart-firefighting technologies and paving the way for future emergency-response tactics in a smart city.

A fan-shaped plasma reactor for mixing enhancement in a closed chamber

This paper introduces a novel fan-shaped plasma reactor, which employs vortex-induced airflow by atmospheric dielectric barrier discharge to enhance mixing and the resulting distribution of the neighboring species of generated ozone. Through stereoscopic PIV and smoke flow visualizations it was demonstrated that mechanisms of suction, vortex creation and ejection of the fluid combine to form a vertical turbulent flow that yields a more controlled and uniform ozone distribution. The performance of the fan reactor was compared to that of a conventional comb reactor for three cross-sectional planes of a space volume simulating the decontamination environment. Results show that with the fan reactor, ozone starts spreading from the center of the plane, which makes the reactor itself responsible for most of the mixing and ozone’s distribution pattern; whereas in the case of the comb reactor, mixing seems to be mostly dependent on the interaction between the reactor’s characteristic flow and the boundaries of the space volume.

Dispersion and upwind-side shear-layer characteristics of forward-inclined transverse jet in crossflow

The effects of the forward inclination of a jet in crossflow on the jet fluid dispersion and upwind-side shear-layer characteristics were studied experimentally in an open-loop wind tunnel. The jet forward-inclination angle θ was varied from 0° to 50°, and the jet-to-crossflow momentum–flux ratio was adjusted between 1 and 2. Short- and long-exposure flow patterns were examined using the laser-assisted smoke flow visualization technique. The binary edge-detection technique was applied to the long-exposure flow images to obtain the jet penetration height. The instantaneous velocities in the upwind-side shear layer of the bent jet were captured by a hot-wire anemometer through a high-speed data acquisition system. The fast Fourier transform method was used to convert the instantaneous velocities from time to the frequency domain. Transverse dispersion of the jet fluids and jet trajectories described by the maximum jet-fluid concentrations were determined using the tracer-gas concentration detection method. The results revealed that at the same jet-to-crossflow momentum–flux ratio, a forward inclination did not significantly influence the flow characteristics and dispersion properties when the inclination angle was smaller than about 12°. At moderate forward inclinations (about 12°–42°), the upstream-side coherent flow structures appeared periodically. The penetration height, transverse width, and dispersion of jet fluids were greater than those of the elevated transverse jet and the slightly forward-inclined transverse jet. At forward inclinations larger than about 42°, the upstream-side coherent flow structures, penetration height, transverse width, and dispersion of jet fluids became the largest when compared with those of the elevated transverse jet as well as the slight and moderate inclination ranges. The jet trajectories in the moderate and large inclination ranges were located at transverse levels higher than those in the slight inclination range and elevated transverse jet due to the appearance of the large upwind-side shear-layer coherent flow structures. The effects of the forward inclinations on the time and length scales of the upwind-side coherent flow structures and the time and length scales of the turbulent eddies were presented and discussed.

Experimental Investigation on the Discharge of Pollutants from Tunnel Fires

Many pollutants are generated during tunnel fires, such as smoke and toxic gases. How to control the smoke generated by tunnel fires was focused on in this paper. A series of experiments were carried out in a 1:10 model tunnel with dimensions of 6.0 m × 1.0 m × 0.7 m. The purpose was to investigate the smoke layer thickness and the heat exhaust coefficient of the tunnel mechanical smoke exhaust mode under longitudinal wind. Ethanol was employed as fuel, and the heat release rates were set to be 10.6 kW, 18.6 kW, and 31.9 kW. The exhaust velocity was 0.32–3.16 m/s, and the longitudinal velocity was 0–0.47 m/s. The temperature profile in the tunnel was measured, and the buoyant flow stratification regime was visualized by a laser sheet. The results showed that the longitudinal ventilation leads to a secondary stratification of the smoke flow. In the ceiling extract tunnel under longitudinal ventilation, considering the research results of the smoke layer height and the heat exhaust coefficient, a better scheme for fire-producing pollutants was that an exhaust velocity of 1.26–2.21 m/s (corresponding to the actual velocity of 4.0–7.0 m/s) should be used. The longitudinal velocity should be 0.16–0.32 m/s (corresponding to the actual velocity of 0.5–1.0 m/s).

Theoretical and experimental studies on the fire-induced smoke flow in naturally ventilated tunnels with large cross-sectional vertical shafts

The study focuses on the fire-induced smoke flow in a naturally ventilated tunnel with large cross-sectional vertical shafts. For the shaft with large cross-section, the smoke is exhausted along four sides of the shaft, in which condition, the complete plug-holing occurs and the fresh air is entrained into the shaft strongly. In this work, studies on the longitudinal smoke temperature rise, smoke mass and heat exhausted as well as air entrainment from the shaft with large cross-section are performed theoretically. A method to predict the maximum smoke temperature rise downstream the shaft and the smoke back-layering length in naturally ventilated tunnel fires is proposed. Meanwhile, a series of small-scale fire tests was carried out by considering heat release rates (HRR) of the fires, shaft heights and adjacent shaft intervals. Experimental results indicate that the smoke back-layering length decreases with the increasing shaft heights, but it seems to be independent of the HRRs that are greater than 25.8 kW employed in the present work. Lastly, the validation of the proposed models is conducted by comparing with experimental results, which indicates that the models predict well the maximum smoke temperature rise on the downstream side of the shaft and the smoke back-layering length for small fires. The present study contributes to a guidance for smoke control in naturally ventilated tunnel fires with vertical shafts.

Experimental study on fire behaviors in narrow bifurcated channel with a confined portal

ABSTRACT A series of experiments were conducted in a bifurcated channel consisting of horizontal main and branch channels to investigate the effects of branch and the confined ratio (the ratio of confined area to cross-section area at branch portal) on fire behaviors. The cross-angle between main and branch channels was 45° and ethanol pool fire fixed in the main channel center was used as fire source. Temperature distribution and flame behavior inside channel were recorded and analyzed. The results show that the branch leads to a much difference in the pattern of smoke flow and makes flame tilt to proximal sidewall to some extent, which is probably ascribed to the asymmetrical air entrainment in spatial. Cold air entrained through branch channel and downstream of main channel is more tend to carry hot smoke into upstream to enhance heat exchange and disturbance to smoke layer, which causes temperature field to asymmetrically distribute and high-temperature region to migrate to upstream. The confined ratio, which determines the proportion of opening area at branch portal, has no obvious influence on temperature distribution of main channel, while it could significantly affect temperature distribution inside the branch channel. The temperature below branch channel ceiling significantly increases with the increase of confined ratio at branch portal, and a corresponding empirical equation is developed. Additionally, under current experimental conditions, the maximum ceiling temperature in bifurcated channel is obviously lower than that of ordinary channel and decreases as confined ratio increases.

A simple correlation for predicting the rise time of fire-induced smoke flow in a tunnel with longitudinal ventilation

The rise time of the smoke flow fronts is an important parameter for fire detection. However, very limited literatures, if not none, have been available to report the predictive correlation or data for the rise time of fire-induced smoke flow in a longitudinally ventilated tunnel. In this study, theoretical analyses were conducted to deduce a simple mathematical correlation considering the effect of longitudinal crosswind disturbance. Computational fluid dynamics (CFD) with large eddy simulation (LES) method was performed to simulate the rising smoke flow with various heat release rates, source-ceiling heights, and wind velocities. The results show that the rise time of smoke flow in the presence of longitudinal crosswind was much larger than that in the absence of longitudinal crosswind, which may be attributed to the enhanced air entrainment that decelerated the rising smoke flow. The numerical results then served as the validation data for the proposed correlation, which confirmed that the smoke flow rise time varied as the −1/2 power of the convective heat release rate, 1/2 power of the ambient wind velocity, and 4/3 power of the height above the fire source. The proposed correlation was found to be capable of predicting the rise time of smoke flow in both windy and windless environments. The findings of this paper contribute to an improved understanding of rising smoke propagation in the tunnel under windy condition. It would also be beneficial to the design of fire detection system for the longitudinally ventilated tunnels.

Critical fan-induced pressure rise for preventing smoke flow multiplicity in longitudinally ventilated sloping tunnel fires: Potential function analysis and numerical simulations

Tunnel fires remain a major concern around the world, especially in sloping tunnels. We investigated the critical conditions for preventing smoke flow multiplicity in longitudinally ventilated sloping tunnel fires and identified multiple flow patterns for smoke movement (i.e., different steady states could be achieved under identical boundary conditions). This results in significant challenges for the design and efficient operation of longitudinal ventilation systems in tunnels. We conducted a potential function analysis and numerical simulations to investigate the critical fan-induced pressure rise for preventing smoke flow multiplicity in a downhill sloping tunnel. Calculation formulas regarding the critical fan-induced pressure rise for preventing smoke flow multiplicity, Pj,max′, were obtained. We demonstrated that a fan-dominated pattern with back-layering flow is physically unstable, and that the fire source location and source strength have an evident impact on smoke flow multiplicity, and on critical values (i.e., Pj,max′andPj,min′) of fan-induced pressure rise. The magnitude of pressure rise derived from critical velocity Pj,cr′ is not sufficient to prevent smoke from spreading upstream in sloping tunnels. Multiple steady states can be achieved if the magnitude of fan-induced pressure rise is between Pj,cr′and Pj,max′. Moreover, Pj,max′ increases more rapidly than Pj,cr′ as the strength of the source increases. The Pj,cr′ is approximately 10–30% less than Pj,max′ when the strength of the source surpasses 20 MW. Therefore, in applications, Pj,max′, rather than Pj,cr′ is essential to ensure the effectiveness of longitudinal ventilation design for smoke control.

Flow and mixing characteristics of a forward-inclined stack-issued jet in crossflow

The flow and mixing characteristics of a forward-inclined stack-issued jet at various inclination angles (θ) and jet-to-crossflow momentum flux ratios (R) were experimentally studied in an open-loop wind tunnel. Flow behaviors were examined using the laser-assisted smoke flow visualization technique. The instantaneous velocities of the upwind-side shear-layer were digitized by a hot-wire anemometer using a high-speed data acquisition system. The instability frequencies in the upwind-side shear-layer vortices were obtained by the fast Fourier transform method. Long-exposure flow images were processed using the binary edge-detection technique to obtain the jet spread width. Transverse dispersion of jet fluids was determined using tracer gas concentration detection. The upwind-side shear-layer vortices revealed four characteristic flow modes: the High impingement-crossflow dominated mode (about θ < 15° and low R), the High impingement-jet flow dominated mode (about θ < 25° and high R), the Low impingement-crossflow dominated mode (about θ > 15° and low R), and the Low impingement-jet flow dominated mode (about θ > 25° and high R). Increasing θ in the crossflow dominated regimes eliminated the upwind-side shear-layer vortices, while increasing θ in the jet flow dominated regimes emphasized the upwind-side shear-layer vortices. Increasing θ at a fixed value of R increased jet spread width in the far field in all modes. In the near field, at x/d < 5 in the High impingement-crossflow dominated regime, the jet spread width was greater than in the Low impingement-crossflow dominated regime. In the jet flow dominated regimes, higher θ values led to greater jet spread width. Transverse dispersion of the jet fluids approached the jet spread width results. In the Low impingement-jet flow dominated regime, transverse dispersion of the jet fluids was significantly increased compared to the other regimes. In addition, the maximum tracer gas concentration was severely reduced at all axial stages, which implied better dispersion of the jet fluids in this regime.

Numerical Investigation of External Wind Effect on Smoke Characteristics in a Stairwell

Fire-induced smoke is the most lethal threat to the residents in high-rise buildings and it is necessary to understand the smoke rising characteristics for the engineering applications of smoke control system. Previous researches only focused on smoke movement driven by buoyancy and stack effect. It is common, however, that external wind can flow inside the building through broken windows and affect the smoke flow. This paper studies the influence of external wind on smoke characteristics in a stairwell. A series of simulations were conducted in a full-scale staircase with top window open. The ambient wind velocity ranged from 0 m/s to 6 m/s and the heat release rate varied from 500 kW to 1500 kW. Results show that the rise time for plume reaching a given height increases with the wind velocity. For a HRR, smoke cannot overcome the wind force when wind velocity exceeds a critical value. Thus a quantitative model is proposed to predict the rise time of plume front considering the hindrance of wind. Moreover, the mass flow rate at the bottom door decreases with wind velocity due to pressure attenuation. However, the CO concentration increases by about 15% with wind velocity, which is a great danger to trapped victims.