FLOOD06499

• County: Statewide
• Basin: Statewide
• Title: Development of Generalized Free Surface Flow Models Using Finite Element Techniques
• Date: 2/1/1979
• Prepared For: US
• Prepared By: COE
• Floodplain - Doc Type: Educational/Technical/Reference Information

FDM model (Fig. 11). Generally, the FEM solution predicts
larger vertical velocity components, perhaps due to the
retention of the vertical momentum equation. Comparison of
the solutions with available field data will be undertaken
once general performance characteristics of the two models
are further defined.
For these steady state simulations, the FDM took about
6 times more CDC 7600 computer time than the FEM. The primary
reason is that, to achieve a steady state solution, the FDM
model must be run through pseudo-time with constant boundary
values until transients from initial conditions die out (about
75-100 days in this case). The FEM model, however, has the
capability of solving the system once with zero time deriva-
tives to arrive at a steady state solution. Comparative costs
for dynamic simulations will depend primarilyupon length of
time step and number of elements used to define the study
region.

SUMI~ARY

The work to date with the horizontal flow model indicates the
following:
(1) Internal continuity errors can be reduced to
acceptable levels by increasing network detail, particularly
in areas of large curvature of the velocity field.
(2) Errors in continuity tend to be reflected more
strongly in the velocity than the depth;
(3) General application of the model to steady state
simulations is feasible at present.
The preliminary work with the vertical flow models
indicates the following:
(1) The finite element method model is less costly than
the finite difference model for steady state solutions.
(2) Simulation of flows in which density gradients are
important requires careful selection of turbulent exchange and
eddy diffusion coefficients.
(3) The finite element model predicts larger vertical
velocities than the finite difference model, perhaps due to
the retention of the vertical momentum equation.
(4) More experience with, and development of, the
vertical models will be required before "production" appli-
cations can be easily made.
Indicated areas of further work are:
(1) Verification of models' perfot'mance when an adequate
data set becomes available.
(2) Development of guidance on selection of turbulent
exchange coefficients, relationship to flow properties etc.
(3) Investigate models' behavior for dynamic simulations.
(4) Evaluate use of stagnation vs. slip boundary con-
ditions in the finite element models.
(5) Extend simulations with the vertical models to
variable breadth problems.

'{