High Fidelity Computations of Grain Silo Explosions and Resulting Blast/Debris Loads on Nearby Structures

Dust explosions are a complex phenomenon that can occur when solid dust particles are dispersed as a cloud. In such configurations, solid particles can burn rapidly as the surrounding air acts as an oxidizer and sustains the combustion of the suspended particles once the reaction is ignited. If the expanding gases generated from the particle combustion are confined due to inadequate venting, the explosion can cause significant destruction to surrounding buildings and personnel. To protect nearby structures from such hazards, a methodology is needed to estimate the resulting airblast and debris. A methodology for predicting the airblast and debris loads generated from accidental dust explosions in grain silos is developed in this paper. The methodology is based on conservative assumptions which provide the initial conditions for coupled computational fluid dynamics (CFD) and computational structural dynamics (CSD) calculations. In the coupled analysis, a large CFD domain encompasses the grain silo CSD domain and a nearby structure. The silo CSD domain is a finite element (FE) model that models the concrete and reinforcing bars which compose the structure. The coupled high-fidelity calculation considers the fluid-structure interaction between the explosive products and the concrete walls during the breakup of the silo. In addition, the reflected airblast pressure is recorded at grid points located on the exterior walls and roof of adjacent buildings. These tracer point loads provide best estimates of reflected pressure and impulse loads on the nearby building as well as estimates of debris size and velocity. The intent of the methodology is to provide conservative load estimates for design purposes. The loads generated from this methodology can then be used for protective construction design and hazard mitigation. Depending on the situation, and as warranted, more refined calculations can be performed to resolve the level of conservativeness.