Abstract: Industrial emissions of small particles can come from the combustion of vehicles, from resuspension due to heavy traffic on roads, from civil construction and also from wind erosion and material handling from storage piles. The last case is emblematic, because, according to emission inventories, diffuse sources of particulate matter may be responsible for an important number of industrial emissions of particulate matter. The largest problem related to the presence of particles in atmospheric air, and of greatest interest to companies, universities, environmental agencies and the population is the risk they can cause to human health. Therefore, estimates obtained through more assertive methodologies are required. The most widely used model for estimating particle emissions from storage piles was developed by the United States Environmental Protection Agency (USEPA). The model, which is simple to use, considers different pile formats, removal and addition of material from them (disturbances), particle size distribution and meteorological conditions, such as wind direction and magnitude. However, it still has some limitations such as (i) the research gap regarding the influence of the percentage of non-erodible particles that make up the pile due to the pavement effect, (ii) limited pile formats, (iii) failure to determine the total mass emitted by resuspension of the material deposited around the piles, (iv) limited wind directions, (v) the difficulty of obtaining the number of disturbances (withdrawal and addition of material) that occur in each area over a period of one year and (vi) dependence on the threshold friction velocity of the materials, a complex variable, still under discussion in the most recent related literature. Regarding the influence of the percentage of non-erodible particles, there is no definitive solution in the related literature. Limitations regarding the pile formats, the calculation of the suspension around the piles and the few wind directions available could be overcome by using the Computational Fluid Dynamis (CFD) technique. In this technique, the computational domain, and the boundary conditions inherent to the yard and local meteorology are fully controllable, allowing the friction velocity distribution to be obtained in the regions of interest. In previous works, it was possible to obtain satisfactory results of wind flow in a full-scale yard, including analysis of the positioning of wind barriers. To diagnose general diffuse emissions at industrial plants, it is necessary operational data from the yards, such as records of disturbance, the amount of material moved, the period of movement and information on the machinery used in operations. The more detailed and organized the information, the greater the chance of successfully diagnosing diffuse emissions. In addition, technical visits to the yards are necessary to assess adverse situations and identify possible sources. Despite all the difficulties concerning input data on emission model, geometry, and mesh definition for CFD are still a major concern when simulating entire and full-scale industrial sites. CFD simulations were performed to identify the Level of Detail for each structure observed in field. It is an important step before the final simulations we intend to perform. The effects of wind direction, obstacles and surface rugosity were also evaluated parameters. Samples of real industrial yards were simulated through CFD to identify the effects of each listed parameter on the erosion and material emission. The results will be relevant when performing the real computational domain.

Keywords: diffuse sources, emission estimation, CFD, computational domain, obstacles, wind direction, level of detail.