Free-Fall Spill of Powder Model

DOE-HDBK-3010-94

4.0 Solids; Powders

5 to 10. At an ARF level of 1E-4 to 1E-5, the variability appears to be in the range of an

order of magnitude or greater.

Figure 4-19 indicates that the RF of the source material is 0.95+ for both powders. The

measured RFs indicate that the airborne particles have not been completely deagglomerated

by the stresses imposed by the event. Powder this fine is extremely atypical of powder

produced by process operations at nonreactor nuclear facilities.

4.4.3.1.3 Free-Fall Spill of Powder Model. Ballinger et al. (January 1988)

proposes a model using the assumption that the powder disperses at a constant angle during

falling and the diameter and velocity of the powder front can be calculated from the angle of

dispersion and properties of the powder. Particles are sheared off during the descent and

remained suspended. The model does not account for the suspension of particles upon

impact. The powder spill model is based upon the following assumptions that appear to be

somewhat inconsistent with observations in subsection 4.4.3.1.1:

the growth rate of the powder front is constant and can be characterized by an

amount of powder airborne is proportional to drag force on the powder; and

the diameter of the powder front at the start of the spill is equal to the

diameter of the container from which it was spilled.

A computer code, PSPILL, was developed to model powder spills. The model was run for

varying values of Mo (mass of powder spilled, kg). An algorithm was developed based upon

the statistical analysis of the results of the computer runs. The algorithm may be used to

predict the ARF if the air density and viscosity are 1.18 kg/m3 and 1.85E-5 Pa-sec,

respectively. The fraction airborne release is:

ARF = 0.1064 (Mo0.125)(H2.37)/

1.02

(4-5)

where:

ARF =

airborne release fraction

Mo =

mass of powder spilled, kg

spill height, m

bulk density of powder, kg/m3.

BP =

In order to determine the bounding ARF, the value calculated from the model must by

multiplied by a factor of 2 (the difference between the average and maximum value for the

Page 4-81