Gordon E. Andrews

Turbulence and turbulent flame propagation—A critical appraisal

Abstract Current theories of turbulence that seem relevant to the structure of turbulent flames are reviewed. The compatibility of such theories with different turbulent flame models is discussed. It is suggested that the turbulent Reynolds number, Rλ, of the reactants is an important controlling parameter in turbulent flame propagation. When Rλ>100, a wrinkled laminar flame structure is unlikely and the turbulent flame propagation is probably associated with small dissipative eddies. It is proposed that the ratio of turbulent burning velocity to laminar burning velocity can be correlated with Rλ.

The relation between polycyclic aromatic compounds in diesel fuels and exhaust particulates

Abstract The polynuclear aromatic compound fractions (PAC) separated by column chromatography from five diesel fuels, a gas oil and sample of kerosene were analysed by capillary column gas chromatography with simultaneous parallel triple detection. The principal polynuclear aromatic hydrocarbons of the fuels are naphthalene, fluorene and phenanthrene and their alkyl derivatives; mutagenic compounds are present in significant concentrations. The principal polynuclear aromatic nitrogen and sulphur compounds of diesel fuel are carbazole and dibenzothiophene and their alkyl derivatives. The PAC of diesel exhaust particulates are similar to those of the fuel, and follow the overall trend of particulate emission with engine load. The 2 to 4-ring PAC in the exhaust are primarily unburnt fuel components. Between 0.2 and 1.0 wt% of these fuel PAC survive the combustion process and comprise a significant concentration of mutagenic compounds in the particulate.

Diesel particulate emissions: The role of unburned fuel

The relationship between diesel fuel composition and that of the solvent organic fraction of diesel particulates was investigated for an old DI Petter engine and a modern DI Perkins engine. Polycyclic aromatic compounds (PAC) were identified using high-resolution capillary column chromatography with a parallel triple detector system for polycyclic aromatic hydrocarbons (PAH), nitrogen-containing PAH, and sulphur-containing PAH. Identification of the PAC using retention indexes was confirmed using an ion trap detector, which was also used to quantify the low-concentration ( < 1 ppm) benzo[a]pyrene. It was conclusively shown for both engines that the bulk of the particulate solvent organic fraction, including the PAH fraction, was unburned fuel. However, there was some evidence that high molecular weight five-ring PAH may have an in-cylinder formation contribution, and it is postulated that this could be due to pyrolysis of lower molecular weight unburned fuel PAH. The contribution of lubricating oil to the particulate PAC is discussed, and evidence is presented that shows the unburned fuel PAC accumulates in the lubricating oil and thus contributes to the particulate PAC via the large lubricating oil component of the particulate PAC.

Real-World Vehicle Exhaust Emissions Monitoring: Review and Critical Discussion

Traffic-related emissions represent a major component of airborne pollution. Historically, dynamometer testing has been most widely used to estimate vehicle emission rates, and these emission rates, in turn, have been used as inputs when modeling traffic-related air quality impacts. However, such conventional drive cycle testing is not considered strictly representative of vehicles under real driving conditions. Therefore, in recent years, significant scientific effort has been focused on the measurement and analysis of real-world vehicle emissions. Here, the use of vehicle emissions monitoring methods (e.g., in-situ methods such as tunnel, inverse dispersion, and remote sensing studies, and in-traffic measures such as probe vehicle and “car chaser” studies) to provide real-world emission estimates is reviewed and discussed in detail. Advantages and disadvantages are identified for the different vehicle emissions monitoring methods, both relative to dynamometer-based approaches and each other. Potential applications of different approaches are also discussed, with particular attention being placed on their complementary use.

The acceleration of flame propagation in a tube by an obstacle

Abstract A quantitative determination was made of the effect of a single baffle of on the characteristics of gas explosions in a 76-mm-diameter closed vessel of large length to diameter ratio ( L D = 21.6 ). Mixtures of methane-air were predominantly used, but other gases were also investigated. Ignition was effected at one end of the vessel. Single hole plates were employed as baffles with varying blockage ratios (20%–80%). The flame speed and rate of pressure rise were greatly enhanced downstream of the baffle. The relative effect of the baffle increased with increasing blockage ratio. It was 8 times more severe with the baffle at 7D from the spark than at 14D. The flow rate of the unburned gas, set in motion ahead of the flame was determined by measuring the pressure drop across the baffle. From the unburned gas velocity the rms turbulent velocity (u′) was determined using experimental correlations of grid-generated turbulence, and from this the turbulent burning velocity and turbulence factor β were calculated. The turbulence factor was found to be equal to the normalized rate of pressure rise. This demonstrated that current turbulent combustion theory (for high Reynolds number flows) can explain and predict the phenomena observed in such combustion regimes. Turbulent flame extinction was predicted for high blockages and experimental evidence of localized flame quenching was found. However, no total flame extinction was observed as the turbulence generated by the baffle was nonuniform and the flame could propagate round local high turbulence regions. The turbulent burning velocity was found to be as high as 110 times the laminar value. In the current literature for vent design a turbulence factor of 10 is suggested for severe cases of turbulence. The present results show the need for reviewing these guidelines if high blockages to the explosion gases exist. A method for estimating β more accurately is tentatively introduced and shown to give very good agreement with the experimental results.

Venting of gas explosion through relief ducts: Interaction between internal and external explosions

Relief ducts fitted to venting openings is a widespread configuration in the industrial practice. The presence of a duct has been reported to severely increase the violence of the vented explosion posing a problem for the proper design of the venting device. Several studies have reported the leading importance--in the whole complex explosion phenomenology--of a secondary explosion in the duct. Modern approaches in the study of simply vented explosions (without ducts) have focused on the study of the interaction between internal and external explosion as a key issue in the mechanisms of pressure generation. The issue is even more relevant when a duct is fitted to the vent due the confined nature of the external explosion. In this work the interaction between internal and external events is experimentally investigated for gas explosions vented through a relief duct. The work has aimed at studying mechanisms underlying the pressure rise of this venting configuration. The study has put the emphasis on the mutual nature of the interaction. A larger scale than laboratory has been investigated allowing drawing results with a greater degree of generality with respect to data so far presented in literature.

Particle size and chemical composition of urban aerosols

The size distribution of airborne particulates (PM10) has been measured by using Andersen Mark II cascade impactors. The measurements were done at four sites, three of which were in the Leeds area, including one roadside site and the fourth at a rural site. On average 10–20% of the mass of urban PM10 particles were found to be below 0.43 μm and 50% below l.5 μm. Extracted samples were analysed for sulphate, nitrate and chloride using ion chromatography, ion selective electrode to determine ammonium and Grans titration to determine acidity. The results show that both sulphate and nitrate peak on the 1.1-μm stage. Nitrate is spread over both coarse and fine modes and is depleted in the finest particles 2.1 μm relative to sulphate. The chloride levels, dominated as they are by sea salt chloride, show a much coarser distribution with <50% being in the fine fractions for either urban or rural areas. The ammonium particulates are totally in the fine particle mode in summer but there is a small amount in the coarse mode in winter. The cumulative size distribution confirms that ammonium is the component with the finest size distribution with 50% <1.0 μm and 80% <1.8 μm. The acidity size distribution is close to the sulphate distribution. The magnitudes of H+ for all sites were very low implying the aerosols are in general well neutralised but the fine particles are more acidic than coarse particles. Rural aerosols are less acidic than urban ones.

The Effect of Ambient Temperature on Cold Start Urban Traffic Emissions for a Real World SI Car

The influence of ambient temperature on exhaust emissions for an instrumented Euro 1 SI car was determined. A real world test cycle was used, based on an urban drive cycle that was similar to the ECE urban drive cycle. It was based on four laps of a street circuit and an emissions sample bag was taken for each lap. The bag for the first lap was for the cold start emissions. An in-vehicle direct exhaust dual bag sampling technique was used to simultaneously collect exhaust samples upstream and downstream of the three-way catalyst (TWC). The cold start tests were conducted over a year, with ambient temperatures ranging from – 2°C to 32°C. The exhaust system was instrumented with thermocouples so that the catalyst light off temperature could be determined. The results showed that CO emissions for the cold start were reduced by a factor of 8 downstream of catalyst when ambient temperature rose from -2°C to 32°C, the corresponding hydrocarbon emissions were reduced by a factor of 4. There was no clear relationship between NOx emissions and ambient temperature. For subsequent laps of the test circuit the reduction of CO and HC emissions as a function of ambient temperature was lower. The time for catalyst light off increased by 50% as the ambient temperature was reduced. The results show that the vehicle used is unlikely to meet the new – 7oC cold start CO emission regulations.

Gas explosions in long closed vessels

Abstract The majority of experimental data on which current vent design practice is based, has been obtained in compact vessels of length lo diameter ratio (L/D) less than 3. This leads to great uncertainty when designing vents for larger L/D vessels. The need for more information on explosions in long enclosures has been noted in recent critical reviews and official guides. In this work methane/air explosions were investigated in closed vessels of large L/D with the aim of providing experimental data relevant to the explosion protection of such vessels. A 76mm diameter tube of an L/D of 21.6 and a 162 mm diameter tube of variable L/D (6.2-18.4) were employed. High rates of pressure rise, associated with fast flame speeds and significant overpressures, characterized the very early stages of these explosions, indicating the need for fast and effective pressure relief within the initial 10% of the total explosion time, Lower flame speeds and lower rates of pressure rise succeeded the fast initial phase. Sma...

Fast flame speeds and rates of pressure rise in the initial period of gas explosions in large L/D cylindrical enclosures

Abstract Flame speeds and rates of pressure rise for gaseous explosions in a 76 mm diameter closed cylindrical vessel of large length to diameter ratio ( L/D = 21.6), were quantitatively investigated. Methane, propane, ethylene and hydrogen mixtures with air were studied across their respective flammability ranges. Ignition was affected at one end of the vessel. Very fast flame speeds corresponding to high rates of pressure rise were measured in the initial 5–10% of the total explosion time. During this period 20–35% of the maximum explosion pressure was produced, and over half of the flame propagation distance was completed. Previous work has concentrated on the later stages of this type of explosion; the development of tulip flames, pressure wave effects and transition to turbulence. The initial fast phase is very important and should dominate considerations in pressure relief vent design for vessels of large L/D .