Industrial Silo Fire and Explosion (Iredell County, Charlotte, NC - December 21, 1997) » Investigation

The investigation as to the cause of the fire and subsequent explosion commenced immediately after the scene was released to the Fire Marshal’s office. The scene was deemed to be unsafe for immediate, detailed, investigation due to the structural damage and the heavy structural material around the top of the silo. This damaged material was removed by a contractor in the presence of the investigating fire marshal and the site was then rendered safe for preliminary investigation efforts. Due to worsening weather conditions, the in-depth investigation was deferred and the effort confined to the perimeter area.

The removal of the remaining product in the silo was required before investigators could enter the silo and proceed with the comprehensive investigation. A contractor used an industrial vacuum and removed the remaining product in the silo on the following Monday. The condition of the remaining sawdust and the burn level in the silo was observed. The scene was photographed and the site was documented by drawings. During the first several days after the explosion, the investigation was hampered adverse weather conditions: ice, snow, and cold temperatures.

The observation indicated a large void area (approximately 12-foot x16 foot x 10 foot) at the bottom of the silo radiating from the screw auger area. One of the "fingers" (Photo #4) of burned material ran to the direction of the access panel that was the focal point of the explosion. The direction of the burned void indicated fire travel in the direction of the screw auger to the compromised areas of the silo shell: the access panels. (Photos #5 and #6). Despite the application of several thousand gallons of water, apparently no water penetrated to sufficient depth to reach the burning wood material.

The fire marshal determined that the fire started when the gear box seized -- burning the belt of the drive unit. The actual fire origin was deemed to have come from the plant side rather than from inside the silo. The day of fire origin was estimated to have been Tuesday or Wednesday. It was surmised that the fire burned void areas out of the sawdust over the auger. The fire also burned from the center toward the bolt holes in the access plates. (Photo #7).

The explosion occurred when the 3'x4' side panel was removed, thus allowing oxygen to enter the burning area. (Photo #8). Because of the compromise of the integrity of the shell by the bolts and access plate, a minute source of oxygen-bearing air permitted the fire to burn in finger-like fashion from the source in the gearbox to the silo wall. This was an area of deep-seated, incomplete combustion, which needed only a source of oxygen to ignite explosively. This is consistent with backdraft-like explosions that have occurred in agricultural product silos. The sawdust also entrained dust to combine with the rapid oxidation of the backdraft to produce the energy sufficient to damage the silo and injure the firefighters. The water applied in the extinguishment attempts apparently either ran off the surface or was absorbed and acted as a tamper over the burned out voids in the wood waste.

Next » Lessons Learned