The addition of suppressants, variation of temperature and variation of pressure also affect the burning velocity.
Flame spread is most rapid when it is directed vertically upward, and an increase in flame spread is directly proportional to increase in the angle of inclination. The reason for this behavior lies in the way in which the physical interaction between the flame and the un-burnt fuel changes as the orientation is varied. Thus while downward spread (-90°) achieves a slow, steady rate of propagation, upward spread (+90°) accelerates as the flame and hot combustion products will preheat the un-burnt fuel directly by convection and radiation. In vertically downward propagation, the flame gases flow away from the un-burnt material and convective transfer does not occur and radiation is unlikely to have an effect as the flame is very small. This is demonstrated in the oxygen index test where vertical downward propagation is involved in the measurement of the burning front of a textile fabric. This test demonstrates a relationship between flame spread and fuel thickness in that the rate of spread is inversely proportional to the thickness of a material. As thickness is increased the rate of spread decreases. The width of a sample has little or no effect on the rate of vertically downward spread.
2.0 Experimentation
At all times during the experimentation standard safety procedures were carried out, these included all hair tied back, lab coats worn at all times in the laboratory and as flammable liquids are being used care was taken to ensure any spillage were cleaned and no unnecessary naked lights present.
2.1 Description of Experimental Apparatus – Flame Stability
The flame stability apparatus consists of a vertical tube burner supplied with metered gas and air through two rotameters. There are various diameter burners and nozzles to be attached to the unit, including the smitthells separation experiment. There is also a horizontal 3m length of 25mm diameter glass tube connected to the mixing block so that it can be filled with a known gas/air mixture.
2.2 Description of Experimental Apparatus – Bomb Calorimeter
Diagram 1. Bomb Calorimeter
The bomb calorimeter consists of a steel bomb made of ausentic steel, which is generally filled with pure oxygen at 25 atmospheres. The high pressure used is to ensure complete combustion of the sample located inside which is no more than 1 gram in weight. The bomb has a screw top for loading and has a total volume of approximately 250ml. The fuel, which is situated in a small crucible suspended from the screw top of the bomb, is ignited by passing a low voltage current through nicrhrome wire which a piece of cotton is tied to, to act as a fuse. The bomb is immersed in about 2 liters of water in a container surrounded by air space and a water jacket. The heat released from the ignition and subsequent combustion of the sample is transferred from the bomb to the water and the temperature rise is a measure of the total heat produced. The bomb calorimeter being used has been slightly modified to allow measurement of the temperature rise by computer. The temperature is measured using a thermocouple, the small voltage is amplified and then fed into an analogue to digital converter, from which the digital information produced is then fed into the pararel port of the computer. The information is then displayed on screen as a temperature reading.
2.3 Description of Experimental Apparatus – The Pensky-Martins Closed Tester
Diagram 2. Penskey-Martens Closed Tester
The Pensky-Martins closed tester consists of a containment vessel, which is placed in an electrically heated metal vessel; the heating power of the metal vessel is 50V. The oil cup is shielded by a stainless steel cover to reduce heat loss by radiation. The oil cup has a closely fitted brass lid through the centre of which passes the stirring mechanism. The lid has four other openings. One is for the thermometer collar and the other three are opened and closed by a slide valve, which is controlled by a spring and lever device rotated by turning the milled control knob. The middle opening is used for introducing the testing flame and the other two for admitting the air required for combustion. The gas test flame is lowered into the containment vessel by the operation of a slide valve. The apparatus may be cooled by connecting a low-pressure air supply to the cooling air nozzle, at the back of the apparatus.
2.4 Description of Experimental Apparatus – The Critical Oxygen Index Equipment
Diagram 3. Critical Oxygen Index Apparatus
The critical oxygen index equipment consists of a gas assembly and valves, oxygen measuring system and a test column and sample holder. There are various controls situated on the front of the instrument that controls various aspects of the apparatus, these are,
- On/Off lever for the oxygen gas line,
- On/Off lever for the nitrogen gas line,
- Fine control needle valve for oxygen,
- Fine control needle valve for nitrogen,
- Main indicating flow meter, scaled 0-25 litres/min for air,
- Flow meter for by-pass flow to oxygen analyser scaled 0-100 ml/min for air.
The right hand side of the equipment houses the gas lines. Two gas inlets for oxygen and nitrogen at the rear of the instrument lead to two filters, from this each pipe leads to a control needle valve fitted with a vernier knob graduated from 0 to 20. This enables a precise setting of the needle valve to be made. An On/Off lever is situated adjacent to the chamber to give a homogenous combined gas flow. Around 99% of this gas flow passes directly to the main indicating flow meter which is calibrated to read from 0 – 25 litre/min (for air). From the flow meter the gas passes directly to the base of the test column. The small by-pass flow (around 1%) goes directly to the oxygen analyser to give a readout of the % oxygen in the gas mixture.
From the oxygen measuring system gas flows to a test column, sample holder and burner. The assembly is housed within a heat resistant Pyrex glass column with a restricted opening. The glass column rests in a casting, which is designed with special entry ports so that the entering gas gives a laminar flow up the column. At the base of the column is a perforated metal screen to catch any fragments or burning droplets, which can come from the sample. The standard sample holder is capable of taking a textile sample of 150mm x 150mm.
The side arm of the textile sample holder is graduated in centre meters so that measurements of the burning rate may be estimated or measured accurately using two inserted thermocouples set at a distance.
3.0 Experimental Method
3.1.1 Flame Stability – Investigation of the Stability Limits and Flame Speed of Gaseous Fuel
A vertical gas burner is supplied with metered gas and air, the flow rate of which, can be altered using two rotamneters. Various different sizes of vertical burners can be attached to this equipment, and qualitative description of the flame was noted at various air/gas ratios. The description of the flame included the change from yellow tipping to non luminous, formation of blue cone, and either lift off or flashback down the burner tube. The height of the blue cone was also measured if it was visible at each reading. The test was repeated several times for varying sizes of burner tube, of which the internal diameter was measures using a micrometer.
3.1.2 Flame Stability – Smitthells Flame Separation
Using the vertical gas burner supplied with metered gas and air, of which the flow rate is regulated by two rotameters. A brass ring was slipped of the mixing tubes and then a glass tube over the ring. The air/gas mixture was ignited from the vertical tube, the glass ring was then lifted above the vertical tube burner after a blue cone had been established, showing that the blue cone should remain above the glass tube not the vertical burner.
3.1.3 Flame Stability – Measurement of Flame Speed
The gas and air supplies were connected to the horizontal flame speed trap. After the air/gas ratios were adjusted through the two rotameters the mixture was ignited, and a time measurement was taken recording the time taken for the flame to traverse the measured length of the tube (2m.) To attain these results the air/gas source was turned off at the time of ignition. The flame speed was determined for various air gas ratios determined by the rotameters.
3.2.1 Bomb Calorimeter – Test 1
The water equivalence of the bomb calorimeter was determined in test 1. This was achieved by using a known weight of benzoic acid. About 0.7 - 0.9 g of benzoic acid and about 12 cm of cotton was weighed out using an analytical balance. The cotton was to act as a fuse inside the bomb calorimeter. A piece of nichrome wire was stretched across the two terminals to act as a short fuse. The cotton was then tied to this ensuring it touches the fuel in the crucible (the benzoic acid.) 10 ml of water was then placed in the base of the bomb the crucible and cradle were then placed into the bomb which was then closed with the screw cover and tightened down. 25 atmospheres of oxygen was then placed inside the bomb, the seals were also checked at this point to ensure that it was air tight. Ensuring the thermocouples were in position just enough water was added to cover the bomb, a record of how much water was used was made. The terminals were then connected, the lid fitted and the stirrer started. The thermocouples were then placed in position ensuring that they did not touch the bottom, after allowing the unit to acclimatize the logging of the temperatures from the thermocouples was started. After five minutes of logging the bomb was ignited and the temperatures logged for a further 20 minutes. After the 20 minutes had passed the bomb was removed from the bomb calorimeter and the pressure was released slowly from it, after dismantling the presence of any residue was noted.
3.2.2 Bomb Calorimeter – Test 2
The procedure from test 1 was repeated using a known weight of Perspex plastic shaving
3.3 Flash Point – The Pensky-Martins Closed Tester.
The containment vessel containing the sample was cleaned before and after each test, after this, the oil cup in the pensky-martins closed tester was filled with the sample to be tested. The lid was then replaced and sample was heated rapidly until the temperature was 25oC below the expected flash point at which point the rate of heating was slowed to 5oC per min. Every 2oC rise in temperature flashing was tested for, by lowering the flame for 1 second into the testing cup, ensuring that the sample is well stirred during the testing period. The first slight explosion due to vapor ignition is observed, that is the temperature which is recorded as the flash point for that sample. After the test the sample was disposed of into the waste can, the apparatus was then cooled using the low pressure air supply and the procedure repeated for each sample. The samples tested by this means were 4-Hydroxy-4-Methyl-2-Pentanon 99%, Cyclohexanone 99.8% and an unknown hydrocarbon liquid.
3.4 Oxygen Index Test – The Critical Oxygen Index Equipment
The equipment was checked and calibrated before use. The sample was cut to the required size and the dimensions and weight recorded, then placed in the sample holder. After removing the chimney attached to the apparatus the sample holder was placed in position and the chimney replaced over the sample. Both switches 2.1 and 2.2 were turned on and 2.3 and 2.4 were adjusted until the total flow was approximately 18 liters per minute and the concentration of O2 and N2 was adjusted until it was reading 25 %, the sample was then purged for 30 seconds before igniting. After ignition the 2.3 and 2.4 were adjusted until the burning front was just maintained and the oxygen percentage recorded. After completion of the run 2.1 and 2.2 were switched off immediately and the chimney removed. Observations of the remnants of the sample were recorded and any way that it burnt. The procedure was repeated for the 3 samples. The samples tested were Polyester, Wool Mix and Calico Med.
4.0 Experimental Results and Calculations
4.1.1Flame Stability – Experimental Results for the Vertical Tube Burner
The results obtained and calculated during the experimentation were as follows, figure 1.
Figure 1. Table of results obtained from Flame Stability Tests
A fuidge diagram was calculated from the results to determine the stoichiometric ratio of a natural gas-air flame in a tube. An example of a fuide diagram is shown in figure 2.
Figure 2. Fuidge diagram for the Natural Gas-Air Flame
4.1.2 Flame Stability – Experimental Results for the Horizontal Burner
The results obtained during the experimentation were as follows,
Figure 3. Results obtained from flame stability tests, horizontal burner.
Flame speed was calculated using the equation shown below,
An example of how flame speed was calculated is shown below, the data used in the example is taken from the results obtained for the air/gas ratio of 10:1,
A graph was drawn of the results obtained and is shown below in figure 4
Figure 4. Graph showing the variation of the flame speed of a natural gas-air flame in a tube with air/fuel ratio.
The graph above shows the variation in Flame Speed of Natural gas-air flame in a tube with air/fuel ratio, and shows the flame speed peaking at around 0.68 ms-1, at an air fuel ratio of approximately 9.8.
4.2 Bomb Calorimetery – Experimental Results
The water equivalent is determined using,
Where
MF = Mass of Fuel (g)
CVFUEL = Calorific Value of the Fuel (cal.g-1)
MWATER = Mass of Water used in calorimeter (g)
WE = Water equivalent of water (g)
TR = Temperature Rise (oC)
MCOTTON = Mass of Cotton (g)
Note: cal.g-1 = x 0.004186 Kj.kg-1
Which rearranges to give
Experimental Data for Benzoic Acid
MF (Benzoic Acid)
Mass of crucible = 5.5991 g
Mass of crucible and benzoic acid = 6.4444 g
Mass of Benzoic Acid = 0.8453 g
CVFUEL = 6.321 k.cal.g-1
= 6321 cal.g-1
MWATER = 1210 g
Temperature Rise TR = End Temp – Start Temp
= 17.66741 – 14.39031
= 3.27 oC
MCOTTON = 0.0030 g
Using this data obtained from the bomb calorimeter the Water Equivalence was determined. This was achieved below
Experimental Data for Perspex
MF (Perspex)
Mass of crucible = 5.0843 g
Mass of crucible and benzoic acid = 6.0876 g
Mass of Benzoic Acid = 1.0033 g
WE = 427.78 g
MWATER = 1225 g
Temperature Rise TR = End Temp – Start Temp
= 17.8745 – 15.4641
= 2.41 oC
MCOTTON = 0.0036 g
Using this data obtained from the bomb calorimeter the internal energy of combustion of Perspex was calculated, this is shown below,
This assumes that the pressure change caused by the oxygen consumption and evolved combustion gases is negliqable, meaning the internal energy change equals the enthalpy release from the specimen.
Therefore the enthalpy of Combustion for Perspex is 16.55 kJ.kg-1
The data required to determine the enthalpy of combustion and the internal energy of combustion was extrapolated from figures 5 and 6, shown below,
Figure 5. Temperature Time graph for Benzoic acid
Figure 6. Temperature Time graph for Perspex
4.3 Flash Point – Experimental Results for the Pensky-Martins Closed Tester
The results obtained from the Pensky-Martins Closed tester for flash point are shown below,
All the flash points obtained are to the nearest 0.5 oC
Figure 7. Results obtained from Pensky-Martins Closed tester
4.4 Oxygen Index – Experimental Results for the Critical Oxygen Index
The results obtained from The Critical oxygen Index apparatus were as follows;
Figure 8. Results obtained from critical oxygen index experimentation
The oxygen index (n) was calculated using the formula shown below,
Where N2 = Volume concentration of Nitrogen
O2 = Volume Concentration of Oxygen
An example of this calculation using the data obtained for Calico med is shown below,
The calculation of Gas Flow was calculated using the equation shown below,
Radius of test column = 4.75 cm
Recommended Total Gas Flow = 18 litres per min = 300 ml per sec
Therefore gas flow up test column is,
5.0 Graphical Analysis and Discussion of Results
The measurements taken during experimentation may have contained random or systematic errors. Repeated results would have improved accuracy for random errors. Possible systematic errors were:
- Human error
- Instrumental error
- Unsuitable method
- Interference in sample
- Sample history
Within the experiment, human errors were likely to be minimal as procedures were carried out carefully. Human error can generally be reduced by reading instruments carefully, and double-checking measurements. Instrumental error may have occurred through incorrect calibration of the apparatus, but was unlikely as the instruments were checked before use. Errors from an unsuitable method were unlikely to occur as it had been carried out before and was shown to be reliable. Interference in the sample may have occurred as a result of dirty apparatus. This could be reduced by properly rinsing and drying sample holders before use. Errors relating to sample history may have arisen if the sample had been incorrectly stored. This error can be reduced by careful storage of the sample in suitable containers. Another source of error could have arisen from the graphs.
5.1Flame Stability
The maximum flame speed for a natural gas-air flame in a horizontal tube was calculated to be 0.68ms-1. This is graphically shown in figure 4. The optimum air fuel ratio determined for this speed was 10:1. No comparisons between the stoichiometric ratios found between the vertical and a horizontal burner was made, because the calculation for the vertical flame speed was not carried out. There are several errors that could be the cause of the varied data obtained from both experiments.
In the measurement of the flow rates of the air and fuel, there was a fluctuation observed in the float meter under certain conditions, making accurate readings very hard. However, more importantly, the conditions at which the three phenomena occur are very subjective. Particularly, it is hard to distinguish a clear cut-off between a yellow-tipped flame and a fully developed one, as variations in the laboratory air will create local regions of yellow flame. Also, a gentle gust will produce premature lift off or light back. As both these conditions are unstable, they are hard to reproduce exactly. It would have been better experimentally to have conducted the experiment under more controlled conditions. Particularly, it required shielding from gusts of air from outside.
The graphical analysis carried out in Figure 4, may produce error as the trend-line that is generated by the computer program may not be accurate, further experimentation would need to be carried out to obtain a wider range of values.
The timing may also be a cause of error, this may be due to the fact that it was timed by a human, which can cause inaccurate timings, if the experimentation was to be repeated, another method of timing the flame speed would be adopted to ensure accuracy of results.
The graph showing the variation of the flame speed of a natural gas-air flame in a tube with air/fuel ratio followed the trends that were expected but deviated at some points.
5.2 Bomb Calorimetery
The water equivalent was determined firstly, in order to calculate the internal combustion energy for Perspex, which then could be converted an enthalpy of combustion. This was achieved by extrapolating the data obtained from the bomb calorimeter thermocouples, shown graphically in figures 5 and 6, to determine the temperature rise. The experimental data was then substituted into the equation shown below to find the water equivalence,
From the calculations the water equivalence was determined to be 427.78g, this calculated result was substituted in the following equation to then determine the internal energy of combustion of Perspex.
From the formula the internal energy of combustion was determined to be 16.55 Kj.kg-1. Assuming the pressure change caused by the oxygen consumption and evolved combustion gases is negligible, the internal energy of combustion equals the enthalpy release from the specimen. Errors may occur, as the total system may not be adiabatic as some heat may be lost from the system, for example from the openings for the thermocouples. The seal which ensures that the bomb is pressure tight may be not fully sealed, which cause gas to be lost and there the assumption that the system is adiabatic can not be made, so the enthalpy release may be slightly different from the internal energy of combustion.
The graphs for benzoic acid and Perspex (figures 5 and 6) show that the jacket temperature of the bomb remained constant. A temperature rise was noted around 37100 secs and the temperature rose by approximately 3°C. The temperature rise peaked at around 37200 secs, then leveled off and continued at this temperature during the rest of the run. Graphical error is unlikely as the computer was connected to the calorimeter and directly recorded the data.
5.3 Flash Point
After test for the flash point using a Pensky-Martins closed tester the flash points of 3 hydrocarbon liquids were determined.
-
Cyclohexanone 99% - 99.5 oC
-
4-Hydroxy-4-Methyl-2-Pentanon 99% - 72.0 oC
-
Unknown Hydrocarbon - 70.5 oC
The flash points are accurate to 0.5 oC.
The text book flash points for the two known chemicals above for a closed cup system is 44.0 oC and 56.0 oC 1 for Cyclohexanone and 4-Hydroxy-4-Methyl-2-Pentanon Respectively. It is clear that there is a great difference in the expected flash points to the actual experimental data. There are several reasons why this may occur for the selected compounds.
Analysing data from flash point manuals indicate that a hydrocarbon with a flash point of 70.5 oC, could be paraffin.
5.4 Oxygen Index
The oxygen index for the various materials, calculated from the experimental results were found to be 16.67, 27.77 and 26.6 for Calico Med, Polyester and wool mix respectively.
The oxygen index (n) was calculated using the formula shown below,
The wool mix had the largest oxygen index, which meant that the fabric required the largest volume concentration of oxygen to support combustion of the material. Calico Med had the lowest oxygen index, which meant that the fabric required less volume concentration of oxygen to support and combustion. It also had a fairly fast burn rate. It can therefore be concluded that this material is more flammable than the others tested in the sense that it would combust in fairly low oxygen enriched environment.
It is clear that the rate of burning is strongly dependent on oxygen concentration, however, this increases the level of radiation, which intensifies the flame making it hotter, which therefore, increases the burning rate. The polyester had the fastest burning rate of 37 seconds, however, it required quite a high volume concentration of oxygen to support combustion and maintain a burning front and therefore, had a high oxygen index. The flame tended to self-extinguish soon after combustion, this may have been due to the components of the material. Because of this it was difficult to maintain a steady burning front during experimentation. Error may have occurred as a result of this and the oxygen index measured may be inaccurate.
The burning behaviour observed during experimentation of the samples was the most notable for the wool mix. The fabric burnt slowly, leaving charred material behind as the burning front progressed. Large amounts of smoke were also produced, this was most likely due to the charred residue formed.
A small amount of charring and dripping of the material was observed for the polyester, and minimal charring was noted for the Calico med. Materials that char on heating such as the test fabrics, build up a layer of char on the surface of the material, which shields the unaffected fuel beneath. This was notably observed for the wool mix and the polyester.
6.0 Conclusions
The aims and objectives of the report as stated at the beginning of the report were as follows:
- To examine the dynamics of fire and explosion using calorimetric methods.
- To quantitatively analyze the measurements obtained from calorimetric instruments.
- To interpret the results obtained in each experiment using graphical analysis.
The first aim was achieved by carrying out the four experiments using calorimetric methods. Behaviour of materials under combustible conditions, flash points using closed tester apparatus, enthalpy of combustion of a specimen in a bomb, flame stability and speed were examined.
The second aim was achieved by presenting the results obtained in a quantitative fashion and interpretating the data using graphical representation and calculations, as shown in the section 3.0 (pg12).
The third aim was achieved by discussing the graphical trends and explaining their significance, looking in particular at possible sources of error.
References
- http://ptcl.chem.ox.ac.uk/MSDS - The Physical and Theoretical Chemistry Laboratory Oxford University
- Centre for Combustion and Energy – Leeds University, LS7 2LR
- Drysdale, D. (1994) An Introduction to Fire Dynamics. New York: John Whiley & Sons
-
mas, P. H. (1981) “Testing products and materials for their contribution for flashover in rooms.” Fire and Materials, 5, 103-111