Journal of Hydroinformatics 4 (2002) 53-62

Fuzzy rule-based model for contaminant transport in a natural river channel

Helen Kettle, Barry Hankin and Keith Beven

Department of Geology and Geophysics, University of Edinburgh, Kings Buildings, Edinburgh EH9 3JW, Scotland, UK Tel: (+44) 0131 650 8511 E-mail: Helen.Kettle@glg.ed.ac.uk

Environmental Science, Lancaster University, Lancaster LA1 4YQ, UK Tel: (+44) 01524 593975 Fax: (+44) 01524 593985 E-mail: b.hankin@lancaster.ac.uk

Environmental Science, Lancaster University, Lancaster LA1 4YQ, UK Tel: (+44) 01524 593892 Fax: (+44) 01524 593985 E-mail: K.Beven@lancaster.ac.uk

Fuzzy rules are used to model solute dispersion in a river dead zone, such that the turbulent diffusion is determined by a fuzzy inference system which relates the local mean velocity shear to the longitudinal velocity fluctuations. A finite-volume hybrid scheme is applied to a non-orthogonal grid for which a mean velocity field is produced using the computational fluid dynamics (CFD) package Telemac 2D. At each cell face fuzzy rules predict a fuzzy number, and these numbers reflect the possible magnitudes of turbulent velocity fluctuations. These are input to the finite-volume model using a single-value simulation method. Multiple model runs produce a fuzzy number for the solute concentration in each cell. The results of the fuzzy model are then compared with data collected in a field experiment with rhodamine dye in the River Severn.

Full article (PDF Format)