Assessment of ignitability 5 19 Figure 7.36 ignition energy of dust clouds Three different electric spark discharge circuits used for determining the minimum ing the discharge with the dust cloud, may be appreciable. Sophisticated elements such as thyrathrons have been employed to solve this problem. However, synchronization of spark and dust cloud can also be accomplished by incorporating a third, auxiliary spark electrode in the spark gap configuration. By discharging just a very small energy in the gap between one of the main electrodes and the auxiliary electrode, the main discharge is initiated. This method was used with success by Franke (1978). Mechanical synchronization constitutes a further possibility. Prior to the experiment the capacitor is then charged to the high voltage required with the spark gap sufficiently long for breakdown to be impossible at that voltage. Pneumatically or spring-driven displace- ment of one of the spark electrodes towards a shorter spark gap, allowing spark-over, is synchronized with the occurrence of the transient dust cloud, for example via solenoids. Boyle and Llewellyn (1950) were probably amongst the first to use the electrode 520 Dust Explosions in the Process Industries displacement method. Its drawback is that the actual spark gap distance at the moment of the discharge is not known. One way of avoiding the synchronization problem is to work with a semi-stationary dust cloud and charge the high-voltage capacitor slowly until breakdown occurs naturally at the fixed spark gap distance chosen. Because of arbitrary variations, the actual voltage at breakdown will differ from trial to trial, and must be recorded for each experiment for obtaining the actual given spark energy 1/2 CV. Figure 7.36 (b) illustrates two versions of the direct high-voltage discharge circuit, one without and one with a significant series inductivity, of the order of 1 mH. This difference can be significant with respect to the igniting power of sparks of similar energies. The induction coil makes the spark more effective as an ignition source by increasing the discharge duration of the spark. Such an induction coil is automatically integrated both in the original US Bureau of Mines circuit, and also in the CMI circuit, as shown in Figure 7.36 (a) and (c). (See Chapter 5 for further details concerning the influence of the spark discharge duration.) If the test is to simulate a direct electrostatic discharge of an accidentally charged non-earthed electrically conducting object, the use of a discharge circuit with low inductance (left of Figure 7.36 (b)) seems most appropriate. 7.10.2.3 The CMI discharge circuit The method for synchronization of dust cloud and spark discharge, which was developed by CMI (see Eckhoff (1975a)), is illustrated in Figure 7.36 (c). The method is similar to the 3-electrode technique in the sense that an auxiliary spark discharge is employed for breaking the spark gap down, but the use of a third electrode is avoided. The energy of the auxiliary spark is about 1-2 mJ. The CMI method requires that the spark energy be measured directly, in terms of the time integral of the electrical power dissipated in the spark gap. Figure 7.37 shows the traces of voltage and current for a spark of net electrical energy 13 mJ, produced by the CMI circuit. The spark discharge was completed after about 280 ps. The general apparatus used by CMI was as otherwise shown in Figure 7.34, i.e. similar to that originally developed by US Bureau of Mines. 7.1 0.2.4 A new international standard method As a part of its efforts to standardize safe design of electrical apparatus in explosible atmospheres, the International Electrotechnical Commission (1989) is considering a new test method for the minimum ignition energy of dust clouds. The draft is to a large extent based on work conducted by an international European working group and summarized by Berthold (1987). The detailed design of the apparatus to be used in a possible IEC test method, in terms of explosion vessel, dust dispersion system, synchronization method, etc. was not specified, but some suitable apparatus were mentioned, including direct high-voltage discharge circuits as well as the CMI circuit. However, no matter which apparatus is chosen, the spark generating system must satisfy the following requirements: Assessment of ignitability 52 1 0 Inductance of discharge circuit 2 1 mH. 0 Ohmic resistance of discharge circuit < 5 R. Electrode material: stainless steel, brass, copper or tungsten. 0 Electrode diameter: 2.0 mm. 0 Electrode gap: 6 mm. 0 Capacitors: low-inductance type, resistant to surge currents. 0 Capacitance of electrode arrangement: as low as possible. 0 Insulation resistance between electrodes: sufficiently high to prevent significant leakage currents. Figure 7.37 Spark gap voltage and spark current versus time during discharge of a 13 ml electric spark from the CMI spark generator. Spark dischar- ge duration 280 p. Energy of trigger spark (spike to the far left) is about 1-2 m/ It will be necessary to take account of the possible influences of dust concentration, dust cloud turbulence and degree of dust dispersion on the test result. Preliminary tests must be carried out to adjust the dust dispersion conditions and the ignition delay such that prescribed minimum ignition energies are actually measured for three specified reference dusts. 522 Dust Explosions in the Process Industries Starting with a value of spark energy that will reliably cause ignition of a given concentration of the dust being tested, the dust concentration being itself a variable, the test energy is successively halved until no ignition occurs in 10 successive tests. The minimum ignition energy is defined to lie between the highest energy at which ignition fails to occur in at least ten successive attempts to ignite the dustlair mixture, and the lowest energy at which ignition occurs within ten successive attempts. 7.1 1 SENSITIVITY OF DUST LAYERS TO MECHANICAL IMPACT AND FRICTION 7.1 1.1 THE INDUSTRIAL SITUATION This hazard primarily applies to powders and dusts with explosive properties, Le. which are able to react or decompose exothermally without oxygen supply from the air. Strong oxothermal reactions may be initiated in layers of such materials if they are exposed to high mechanical stresses and fast heating by impact or rubbing, either accidentally or as part of an industrial process. 7.1 1.2 LABORATORY TESTS 7.1 1.2.1 Drop hammer tests As summarized by Racke (1989), a number of impactlfnction sensitivity test methods have been developed in several European countries, as well as in USA and Japan. The most common design concept for the impact test is the drop hammer, as illustrated in Figure 7.38. Verein deutscher Ingenieure (1988) also mentioned the very similar test by Lutolf (1978) as a suitable standard method. In the Lutolf test the dust sample size is about 0.10 g and the theoretical maximum drop hammer impact energy 39 J (5 kg, 0.8 m). Up to ten trials are conducted and observations are made with respect to occurrence of explosion, flame, smoke or sparks. If all ten tests are negative, a new test series is conducted with the dust samples wrapped in thin aluminium foil (10 pm thickness), in case the aluminium should have a sensitizing effect on a possible exothermal reaction. If the tests with aluminium are positive, a new test series without aluminium is conducted. The American Society for Testing and Materials (1988) adopted the US Bureau of Mines drop hammer method as their standard. Using a fixed drop hammer weight (2.0 or 3.0 kg), the drop height H,, giving 50% probability of a positive reaction, is determined. The lower H,,, the more sensitive the material is to impact ignition. In the test description Assessment of ignitability 523 Figure 7.38 5 kg and height of fall 7 m (From Verein deutscher Ingenieure, 1988) Drop hammer test for dust layers by Koenen, Ide and Swart (7 96 I). Drop hammer mass it is emphasized that the observation of the reaction of the sample is one of the difficult points in impact sensitivity testing. A positive test result is defined as an impact that produces one or more of the following phenomena: (a) audible reaction, (b) flame or visible light, (c) definite evidence of smoke (not to be confused with a dust cloud of dispersed sample), and (d) definite evidence of discolouration of the sample due to decomposition. The problem arises with reactions that yield no distinguishable audible response, no flame, and little sample consumption. The decision concerning reactiodno reaction in these cases must be based primarily upon the appearance of the sample after the test. The impact in most cases will compress the sample into a thin disc, portions of which may adhere to the striking tool surface, the anvil, or both. One should then inspect the tool and anvil surfaces and look for voids in the powder disc and discolouration due to decomposition in areas where voids occur. If there is discolouration from decomposition, the test trial is to be considered as positive. If there are small voids but no discolouration, the trial should be regarded as negative. In the case of doubt as to whether or not discolouration is present, the trial is to be regarded as negative. If the only evidence is a slight odour or a small amount of smoke, which may be a dust cloud from dispersed sample, the trial should also be considered negative. 7.1 1.2.2 Friction tests As pointed out by Racke (1989), several different friction tests have been devised, including three described by Gibson and Harper (1981). One of these is illustrated in Figure 7.39. 524 Dust Explosions in the Process Industries Figure 7.39 Example of laboratory method for testing the sensitivity of powders to mechanical rubbing/friction (From Gibson and Harper, 1981) 7.1 2 SENSITIVITY OF DUST CLOUDS TO IGNITION BY METAL SPARKS/HOT SPOTS OR THERMITE FLASHES FROM ACCIDENTAL MECHANICAL IMPACT 7.1 2.1 THE INDUSTRIAL SITUATION Dense clouds of metal sparks, and also hot surfaces, are easily generated in grinding and cutting operations. Such operations are therefore generally to be considered as hot work, which should not be permitted in the presence of ignitable dusts or powders. However, the evaluation of the ignition hazard to be associated with accidental impacts is less straight-forward. Such impacts can occur due to mis-alignment of moving parts in powder processing equipment, for example in grinders and bucket elevators. Or foreign bodies such as stones and tramp metal can get into the process line. Whether or not metal sparks/hot spots or thermite flashes from single accidental impacts between solid bodies, can in fact initiate dust explosions, has remained a controversial issue for a long time. It now seems that in the past ‘friction sparks’ have been claimed to be the ignition sources of dust explosions more often than one would consider as reasonable on the basis of more recent evidence. However, as long as necessary conditions for such impacts to be capable of initiating dust explosions have been unidentified, one has been forced to maintain the hypothesis that such sparks may be hazardous in general. This in turn has forced industry to take precautions that may have been superfluous, and caused fear that may have been unnecessary. Generation of metal sparkshot spots by accidental mechanical impacts is a complex process, involving a number of variables such as: 0 Chemistry and structure of the material of the colliding bodies. 0 Physical and chemical surface properties of the colliding bodies. Shapes of the colliding bodies. 0 Relative velocity of the colliding bodies just before impact. 0 Impact energy (kinetic energy transformed to heat in an impact). Single or repeated impacts? Assessment of ignitability 525 Whether a given dust cloud will be ignited by a given impact not only depends on the specific dust properties, but also on: 0 Dust concentration and dynamic state of the dust cloud. 0 Composition, temperature and pressure of the gas phase. In view of the great number of variables and the lack of an adequate theory, it is clear that the ignition experiments on the basis of which the practical hazard is to be assessed, should resemble the practical impact situation as closely as possible. 7.1 2.2 LABORATORY TESTS No standardized test methods have so far been traced, but the ability of metal sparks/hot spots from grinding and cutting to ignite dust clouds has been demonstrated in laboratory tests by several researchers, including Leuschke and Zehr (1962), Zuzuki et al. (1965), Allen and Calcote (1981) and Ritter (1984). (See Chapter 5.) Laboratory test methods for the incendivity of single accidental mechanical impacts seem to be less numerous. A test apparatus developed by Pedersen and Eckhoff (1987), is illustrated in Figure 7.40. Figure 7.40 Apparatus for determining the sensitivity of dust clouds to ignition by single accidental mechanical impacts (From Pedersen and Eckhoff, 1987) 526 Dust Explosions in the Process Industries The basic principle of impact generation is that a spring-loaded rigid arm, which can swing around a fixed axis, and carries the test object at its tip, is released and hits a test anvil tangentially at a known velocity. Depending on the normal contact force during impact, the peripheral velocity of the tip of the arm will be more or less reduced. By knowing the mass distribution of the arm and the peripheral velocity of its tip just before and just after impact, the impact energy can be estimated in terms of loss of kinetic energy of the arm. The impact force is varied by varying the excess length of the arm compared with the distance from the arm axis to the anvil. Figure 7.41 gives an expanded view of the test object holder at the arm tip. The dust cloud was generated by dispersing a given quantity of dust from a dispersion cup by a short blast of air. The dust concentration of the transient cloud near the point of impact, at the moment of impact, was measured by a calibrated light attenuation probe. (See Figure 1.76 in Chapter 1. ) Figure 7.41 Expanded view of test object holder of apparatus shown in Figure 7.40 (From Pedersen and Eckhoff, 1987) Figure 7.42 shows some typical results from experiments with the apparatus shown in Figure 7.40. Further details of this kind of experiments are discussed in Chapter 5. Because of the lack of generally accepted test methods, it has been suggested that the sensitivity of a dust cloud to ignition by metal sparkshot spots from accidental impacts may be correlated to the sensitivity of ignition by other sources, such as electric sparks. As discussed in Chapter 5, Ritter (1984) found a correlation involving both the minimum electric spark ignition energy and the minimum ignition temperature as determined by the BAM furnace. Table 7.2 indicates a correlation with the minimum electric spark ignition energy alone. Assessment of ignitability 527 Figure 7.42 Frequency of ignition of clouds of dried maize starch in air as a function of impact energy at 16 m/s and 24 m/s peripheral velocity of approach of the arm tip. Bars indicate k 1 standard deviation. Impacts between titanium and rusty steel (thermite flashes) (From Pedersen and Eckhoff, 1987) Table 7.2 Results from single-impact ignition tests of dust clouds of different minimum electric spark ignition energies, using a 20 J thermite flash impact between titanium and rusty steel (From Pedersen and Eckhoff, (1987). 7.1 3 MINIMUM EXPLOSIBLE DUST CONCENTRATION 7.1 3.1 THE INDUSTRIAL SITUATION For a given type of explosible dust, dispersed as a cloud in air, there is a reasonably well defined minimum quantity of dust per unit volume of air below which the dust cloud is not able to propagate a flame. (See Chapter 4 for full discussion.) In theory, therefore, one could eliminate the possibility of dust explosions by ensuring that the dust concentration does not exceed this minimum limit. In practice, however, most process equipment in plants where powders are manufactured and handled will always contain large quantities of powder, and hence this principle of preventing dust explosions is not practicable in general. There are, however, some types of process equipment to which the principle may be adapted in practice (see Section 1.4.3.2). 528 Dust Explosions in the Process Industries One example is dust extraction systems designed for the purpose of extracting a relatively small quantity of fine dust from a coarse main product, as in grain silo plants. In such cases the concentration of dust in the system can often be controlled to some extent by controlling the flow of air. It is then essential, however, that the air velocity is maintained sufficiently high to prevent dust from depositing on walls of ducting etc. , since such deposits, if redispersed, may form clouds of explosible concentration. Another type of equipment that can be protected by keeping the dust concentration sufficiently low, is systems for electrostatic powder painting. In such systems the concentration of particles in the air is relatively uniform and fairly easy to control. In fact, several countries have imposed specific maximum permissible average dust concentrations in the spraying booth, based on estimates of the minimum explosible dust concentration. (See Section 1.5.3.5.) 7.1 3.2 LABORATORY TESTS Experimental determination of the minimum explosible dust concentration is discussed in detail in Section 4.2.6.2 in Chapter 4. This also includes comparisons between various test methods in use. 7.1 3.2.1 Tests developed in USA In the standard test used in USA and UK for a number of years and described by Dorsett et al. (1960), a known quantity of the powder was dispersed as a cloud in a slim, vertical, cylindrical container of 1.2 litre volume and exposed to a continuous spark ignition source. Starting with very small powder quantities and repeating the test with steadily increasing amounts, a critical quantity was reached at which the dust cloud ignited. The critical mass of dust, divided by the volume of the test container, was taken as the minimum explosible dust concentration (MEC). It was felt that the traditional test method was not fully satisfactory. On the one hand, the continuous ignition source was located in the lower part of the vertical, elongated explosion vessel, and this would allow the dust cloud, rising from the dispersion cup of the vessel bottom, to become ignited before having been fully dispersed throughout the entire vessel volume. Hence, the real concentration of dust in the cloud at the moment of ignition was likely to be higher than the nominal concentration estimated by dividing the mass of dust dispersed by the total vessel volume. This error would generally lead to underestimation of MEC. On the other hand, the traditional ignition source was a continuous train of relatively weak electric sparks that may not have been sufficiently energetic for igniting dust clouds of concentrations near the true limit for self-sustained flame propagation. This would generally yield overestimation of MEC. The effects of these two factors tend to cancel each other, and this may be the reason for the surprisingly good agreement that has in some cases been obtained between MEC values from the traditional small-scale lab test, and large-scale experiments. For example, Jacobson er al. (1961) found that various grain dusts and starches all gave MEC’s of the order of 50 s/m3 in the small lab-scale test, which compares favourably with the value 60 g/m3 found for a [...]... with traversing cylinder for measuring local dust concentration in the vicinity of the ignition source (From Nordtest, 198 9) I Figure 7.48 Principle o Nordtest Fire 0 1 1 dust cloud sampling cylinder (b) compared with that o a f f simple cylindrical cup (a) 534 Dust Explosions in the Process Industries As discussed in Section 4.26.2 in Chapter 4, there are indications of this test method yielding unexpectedly... pure air 7.1 9 INFLUENCE OF OXYGEN CONTENT IN OXIDIZING GAS ON THE IGNlTABlLlTY AND EXPLOSlBlLlTY OF DUST CLOUDS 7.1 9. 1 THE INDUSTRIAL SITUATION Full and partial inerting is discussed in Sections 1.3.6 and 1.4.3 in Chapter 1 The possibility of dust explosions in process equipment can in principle be effectively eliminated by substituting the air by a gas which makes flame propagation in the dust cloud... for determining Ksr values according to the I S 0 standard, the dust dispersion system, the ignition source strength and the ignition delay must be tuned in such a way that the products of the maximum rates of 544 Dust Explosions in the Process Industries pressure rise measured and the cube roots of the vessel volumes equal the Ksr values that would have been measured for the same dusts in the 1 m3... The version of the 15 litre vessel used in the second step is shown in Figure 7.47, and the basic principle of the traversing dust sampling cylinder is illustrated in Figure 7.48 7.1 3.2.4 Possible international standard The International Electrotechnical Commission ( 199 0) is evaluating a test method based on the 20 litre Siwek ( 198 8) sphere Nordtest ( 198 9) and the 1 m3 vessel of the International Standardization... useful information for assessing the gain in safety obtained if partial inerting is used The apparatus in Figure 7.64 may be used for measuring maximum explosion pressure and rate of pressure rise as functions of the oxygen content in the atmosphere, which provides further information about the effect gained by partial inerting The inlet for combustible gadair would then instead be used for the mixture... Hartmann tube apparatus for determining the influence of the oxygen content in the atmosphere on the ignitability of dust clouds Perspex cylinder During this process the air that was originally in the Perspex cylinder leaks to the atmosphere The small reservoir is now pressurized with the appropriate gas mixture to a pre-determined level, found in earlier trials to give the best dust dispersion conditions... local dust concentration in the vicinity of the ignition source, at the same instant as the ignition source would be activated in the first step, is determined using the dust mass giving 50% of ignition, and exactly the same dust dispersion method as in the ignition tests The arithmetic mean of five consecutive concentration measurements is taken as the minimum explosible dust concentration The version... 536 Dust Explosions in the Process Industries Figure 7.50 Photograph of the version of the Hartrnann bomb shown in Figure 7. 49 The ignition sources used include continuous trains of electric sparks, single synchronized sparks, synchronized chemical igniters and glowing resistance wire coils Versions of the two latter are shown in Figure 7.51 For determination of maximum pressures the nature of the. .. opening time The container is connected to the explosion chamber via a 19 mm 0 perforated semicircular spray pipe The diameter of the holes in the pipe should be in the range 4-6 mm The number of holes is chosen such that their total cross-sectional area is approximately 300 mm’ 538 Dust Explosions in the Process industries Figure 7.53 Typical set of results from Hartmann bomb test of a given dust. .. impossible Since the use of large quantities of inert gas in a plant can be expensive, it is important to limit the inert gas consumption to the extent possible For most dusts it is not necessary to substitute the entire atmosphere in the actual area by e.g nitrogen, carbon dioxide, or other inert gas to obtain inerting Hence, it is essential to know the critical gas composition for inerting the dust in question . the minimum explosible dust concentration. The version of the 15 litre vessel used in the second step is shown in Figure 7.47, and the basic principle of the traversing dust sampling cylinder. included in the standard specified by the American Society of Testing and Materials ( 198 8a). 536 Dust Explosions in the Process Industries Figure 7.50 Photograph of the version of the. been devised, including three described by Gibson and Harper ( 198 1). One of these is illustrated in Figure 7. 39. 524 Dust Explosions in the Process Industries Figure 7. 39 Example of