Figure "Risk of dust explosions in compressor by synthetic materials": Dust explosions in engines are not as rare as the scarcity of their documentation might lead one to think. Symptoms of dust explosions are often observed in the containment tests required during engine development (a dark cloud emerges from the front of the fan and becomes a reddish fire). If large amounts of titanium dust are present (e.g. after fan blade rubbing), white burning flames can also be observed. This phenomenon can also be observed at the engine outlet in the bypass duct region, where it is caused by the dust moving through the bypass duct (top diagram).
The pressure wave created by the dust explosion can throw fragments of the blading or first guide vane stage along with the housing ring out of the front of the engine (bottom diagram).

Materials that create explosive dusts are often used in the following engine parts (bottom diagram):

• Housings
• Compressor guide vanes
• Compressor runner blades
• Rub coatings

The flammable dust created can be of various different compositions:

• Synthetic abrasions from abradables made from filled polyester and epoxy resins. Fill materials include hollow spheres of phenolic resin, aluminium powder, and graphite.
  
• Abrasions from inorganic abradables, such as sprayed nickel/graphite coatings
• Abrasions from soft metallic rub coatings such as sprayed aluminium coatings
• Dust from fragments of fibre-reinforced synthetics, such as carbon fibre-reinforced epoxy.
• Abrasions from metallic blades made from titanium or aluminium alloys.
• Metallic housing abrasions (magnesium-and aluminium alloys)
• Organic particles used to clean the compressor

Figure "Risky compressor parts for dust explosions": The amount of material whose abrasions could cause a dust explosion in modern tactical aircraft is about 1 kg. This contains a dangerously large amount of energy; roughly the same as that of an equal amount of propane gas by weight.

Potential vital weak points that could be seriously damaged by a dust explosion in an engine are:

• Axial flange and screw connections of compressor housings (1), overstress.
• Radial flange and screw connections of compressor housings (2), overstress.
• Rotor- and stator blades in the compressor due to elastic deformation and contact (3)
• Adjustment guide devices: shutting and/or overstress (4).

Tests in pipes with a diameter of about 1 meter can provide insight as to the stresses a compressor (housing, blading) experiences from a dust explosion (Ref. 9.4-3). In dust explosions in pipe sections sealed at both ends, which correspond to compressor housings with several stages on either side of the explosion, dampening of the explosion sequence in the direction of the sealed pipe end only occurs in relatively unreactive dusts (class St1). In more reactive dusts, the explosion speeds in pipes sealed at both ends are as high as those in pipes sealed at only one end. As with combustible gases, the maximum explosion pressure of combustible dusts in completely sealed pipe sections is several bar above that in pipe sections sealed only at one end. As per Ref. 9.4-3 “…it remains to be summarily proven that in dust explosions in pipe sections, displacement- and therefore flow- and turbulence effects determine the course of the explosion. With this, the K_St value (Fig. "Sequence of influences on dust explosions") that originally determined the strength of explosions in sealed containers loses much of its importance…“

Figure "Danger of Dust explosion in spin-test rigs":In order to determine burst RPM and safe load cycle numbers of engine rotors, a so-called “vacuum testing rig” can be used. It might be better to speak of underpressure-testing rigs, since a high vacuum is not actually necessary, as long as the performance loss due to air friction is minimized. As the excerpt at right from Ref. 9.4-2 shows, there is a possibility of flammable dust being created during trials in these types of testing rigs. If air enters the chamber (for example through a broken observation window, fragments lifting the lid, or when the chamber is flooded with air in order to open the lid) a spark can cause a dust explosion with severe consequences. Therefore it is especially important, that the lid is securely fastened during the trial (the underpressure is not sufficient for this task!) and that, after the trial, the testing rig is flooded with a protective gas appropriate to the expected dust. In the case of titanium abrasions, CO₂ or nitrogen (TiN is created) are not suitable . These gases react with titanium particles as strongly as air does. If in doubt, the noble gas argon should be used.

Figure "Determining the explosion behavior of dust" (Ref. 9.4-5): There are standard testing rigs that are used to determine the explosiveness of dusts. Trials take place in an open (modified Hartmann pipe, top diagram) or in a closed apparatus (bottom left diagram). The dust`s capacity for exploding is proven if ignition of a dust/air mixture is followed by a spreading of flames accompanied by a pressure increase in closed containers.

In the modified Hartmann pipe (a vertical pipe chamber with an openable lid at the top) dust is whirled up from the bottom by a blast of air and then ignited with a spark gap. This corresponds to the conditions in a testing rig when flooded with air (Fig. "Danger of Dust explosion in spin-test rigs"). If the flame spreads, with or without lifting the lid, the mixture is considered to be capable of dust explosions. If the mixture does not ignite, further tests in enclosed apparatus are necessary before it can be classified as incapable of dust explosions.

In a closed testing rig, the dust sample is blown into the 1 m³ test chamber with pressurized air and a set ignition energy is applied (bottom right diagram). The explosion sequence is observed and recorded. From the pressure-time curve, the explosion pressure and the pressure increase rate (time) Dp/Dt are deduced (Fig. "Sequence of influences on dust explosions"). If pressure builds up (>0,5 bar) the mixture is considered to be capable of dust explosions (in the tested concentration). This trial is more representative of conditions in engines where abrasions enter the compressor and are ignited.

Figure "Sequence of influences on dust explosions" (Ref. 9.4-5): The specific values given were determined in a sealed testing rig with various mixture ratios (Fig. "Determining the explosion behavior of dust").
The maximum explosion pressure dependant on oxygen percentages (top left diagram, Fig. "Determining the explosion behavior of dust") and the explosion pressure dependant on dust concentration (top right diagram) are determined in individual trials. No dust explosion occurs below the limits (oxygen concentration and lower explosion limit). Therefore, if the amount of abrasions is small compared to the air flow, which is usually the case in normal rubbing, there is no danger of a dust explosion in the compressor. In the case of capital compressor damage with blade failures there will probably be a stall and therefore a low air flow rate with high dust levels. Experience has shown that in this situation, explosive mixtures can be created in parts of the compressor. The maximum pressure increase (by time) dependant on dust concentration also depends on the volume of the container(bottom diagram). From this, a dust- and trial condition-specific characteristic value independent of the container size (K_St-value, compare Fig. "Determining the explosion behavior of dust") can be determined. It is especially important for the stress levels of the engine parts (housings and bladings). Naturally the pressure building in the compressor is determined by the air resistance of the blading in front of and behind the point of explosion and should be assessed accordingly.

Typical abradables with a synthetic matrix for blade tips in the front compressor such as

• graphite-filled epoxy resin
• AlSi powder-filled polyester resin

are capable of dust explosions (class St 1; K_St-value<200 bar m s^-1) and comparable with coal dust, powdered sugar, or air/gas mixtures of propane or methane. Aluminium powder is classified considerably higher (class ST3; K_St-value > 300 bar m s^-1).
The K_St-value of polyester resin filled with 60% by weight AlSi does not differ from that of pure resin. This value is considerably lower than would be expected from the amount of light metal alone. Epoxy resin has a K_St value of about 160 bar m s^-1. The graphite slightly lowers this value.

Figure "Bird strike causing dust explosion" (Example "Dust explosion after bird strike"):After a heavy bird strike, apart from heavy damage to the fan, it was discovered that the axial flange of the booster housing had been torn open due to an overstressing of the screw connection (bottom left diagram). This damage was traced back to a dust explosion in the booster (bottom left diagram) caused by synthetic abrasions from the primary damage.