Determination of Particle Collision Efficiencies for Flocculent Transport Models

Abstract

The object of this study is to investigate the data needs for accurate estimation of particle collision efficiencies ({alpha}). Batch flocculation and vertical flocculent settling studies were conducted on aquatic sediment particles in the size range of 2--80 {mu}m in diameter. Dynamic particle-size distribution data sets were generated using a 256-category binary-collision-based flocculent settling model and subsequently compressed into 128, 64, 32, 16, 8, and 4 logarithmically varying particle-diameter categories. A previously developed parameter-estimation framework was used to estimate {alpha} values from the compressed data sets. Variability in the {alpha} estimates indicates that it is not an artifact of the experimental data. The residual function value after 20 estimation iterations increased with the discretization levels, indicating shallower residual function slopes. Parameter estimate accuracy was unaffected by the levels of data sparsity considered. Increased random noise resulted in consistently lower estimated collision efficiencies. Rate of convergence was unaffected by random noise but decreased with data sparsity levels. Particle water column residence times were significantly overestimated using four category distributions.