Estimation of natural river longitudinal dispersion coefficient based on particle swarm optimization algorithm

The longitudinal dispersion coefficient is very important for studying the migration of river pollutants.At present, scholars have put forward many empirical formulas and theoretical formulas, but the accuracies of the calculation results are not high.In this paper, based on the measured values of the longitudinal dispersion coefficient and the corresponding river hydrological data and river morphology data, through the weighted combination of the dual objectives of the calculation accuracy of the longitudinal dispersion coefficient and the correlation between the calculated value and the measured value, the particle swarm optimization algorithm( a data-driven algorithm) was used to optimize the objective function after the weighted combination, so as to obtain the undetermined coefficients in the general expression of the longitudinal dispersion coefficient.Finally, a new natural river longitudinal dispersion coefficient calculation formula with high accuracy and high correlation was proposed.The verification results showed that the formula can be applied to the longitudinal dispersion coefficient estimation of common natural rivers with longitudinal dispersion coefficient between 0 and 100 m2/s.