Elimination of Time Assignment Bias in Estimates of Dispersion Coefficient

Dispersion coefficients are commonly determined by regressing column effluent data against predictions of transport equations. When column effluent is collected in fractions, measured concentration is an average over the collection interval. A method is presented for estimating dispersion coefficients by regressing column data against the integral of the transport equation. This dispersion coefficient is compared to estimates obtained when effluent concentration is assigned to the end or middle of collection intervals. Dispersion coefficient estimates obtained by regressing effluent concentration against the integral of the transport equation were as much as 80% greater than those obtained by assigning effluent concentration to the end of the collection interval, and within approximately 20% of those obtained by assigning effluent data to the middle of the interval. The amount of difference depended on model variables of collection interval and Peclet number.