Print Email Facebook Twitter Implementation of a wetting and drying algorithm in a finite element model Title Implementation of a wetting and drying algorithm in a finite element model Author Kroon, J. Contributor Labeur, R.J. (mentor) Vuik, C. (mentor) Talstra, H. (mentor) Stelling, G.S. (mentor) Faculty Civil Engineering and Geosciences Department Hydraulic Engineering Date 2009-09-21 Abstract In numerical modelling of water movement wetting and drying is a well known problem. The governing equations are not valid in the dry part of the computational domain which may result in problems with mass conservation, negative water depths and artificially enlarged gradients. A method is proposed that allows for the surface elevation to become negative while strict positivity is demanded from the water depth. In this way mass conservation is guaranteed. It is investigated whether this procedure is mass conservative, efficient and robust. To this end a simple finite element discretization of the inviscid shallow water equations (SWE) is derived. The resulting procedure is validated with several one- and two-dimensional analytical solutions for: 1) a one dimensional dam break, 2) flow over a long crested weir, 3) a one-dimensional oscillating water surface in a parabolic basin, 4) the runup of long waves on a beach, 4) a two-dimensional standing wave in a parabolic basin and 5) the spreading of a parabolic flood wave in two dimensions. In addition two laboratory experiments are used for validation. One experiment with solitary wave runup on a conical island and one experiment with a two dimensional dam break. In general the methods performance is good. However in two-dimensions it can be beneficial, in case of small gradients at the wet/dry interface, to use a lumped mass matrix at partially dry elements too. The wetting and drying iteration converges on average in 2 to 3 iteration steps. In some cases bifurcations and mass errors can occur. However mass errors are caused by rounding errors and can be resolved by using double precision. The occurrence of bifurcations is much less frequent in case of calculations in double precision and can be minimized by adjusting the BiCGSTAB settings. Subject wettingdryingfinite elementsdam breaknumerical model To reference this document use: http://resolver.tudelft.nl/uuid:e82ee739-a03e-441f-bd9e-637c0772ac88 Part of collection Student theses Document type master thesis Rights (c) 2009 Kroon, J. Files PDF Finalreport.pdf 3.85 MB Close viewer /islandora/object/uuid:e82ee739-a03e-441f-bd9e-637c0772ac88/datastream/OBJ/view