Premium
Stability Study of a Longitudinal Dispersivity Identification Method
Author(s) -
Pisani Stefano,
Tosi Nicola
Publication year - 1993
Publication title -
groundwater
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.84
H-Index - 94
eISSN - 1745-6584
pISSN - 0017-467X
DOI - 10.1111/j.1745-6584.1993.tb00835.x
Subject(s) - gaussian , isotropy , nonlinear system , standard deviation , porous medium , reliability (semiconductor) , probability density function , approximation error , identification (biology) , mathematics , statistics , physics , porosity , geology , geotechnical engineering , optics , thermodynamics , power (physics) , botany , quantum mechanics , biology
The knowledge of the value of longitudinal dispersivity is very important in the simulation of solute transport. Therefore, a method for the identification of longitudinal dispersivity in a solute transport laboratory experiment was developed. The solute was assumed to be ideal and the medium isotropic and homogeneous, with a one‐dimensional geometry. The sensitivity of the inversion method to a uniform random noise in porosity, which accounts for possible heterogeneities of the medium, and to concentration measurement errors, which are always present in a laboratory experiment, was analyzed. This analysis is necessary to estimate the reliability of the method. Then, supposing the concentration error distributed according to a Gaussian with zero mean, the probability distribution of the relative identification error was computed as a function of the standard deviation of the Gaussian. The mean value of the relative identification error was found to be positive, due to the nonlinearity of the identification algorithm.