Geoscience Reference
In-Depth Information
process parameters are often calibrated using measured data or determined using
empirical formulas.
(3) Model Calibration . The computational model should be calibrated using the avail-
able datameasured in the study reach to insure that the aforementioned parameters
are estimated correctly, that the empirical formulas are chosen appropriately, and
that the observed physical processes are generally well reproduced by the model.
The calibrated model can then be applied to predict the physical processes in
various scenarios.
(4) Interpretation of Simulation Results . Because sediment transport models are
highly empirical and the model development and application processes are not
infallible, engineering judgment should be used in the interpretation of simu-
lation results. Consulting with model developers, senior scientists, and local
engineers who are familiar with the study channel can enhance confidence in
the end results. In addition, efficiently grouping important results using attrac-
tive graphs and tables permits an easy understanding and communication among
model developers, users, and report readers.
(5) Analysis of Uncertainties . Sources of uncertainties include model conceptualiza-
tion, boundary conditions, model parameters, and data. Uncertainties may be
reduced by using a more adequate model, selecting appropriate boundary condi-
tions, calibrating model parameters carefully, and collecting more reliable data.
Sensitivity analysis and stochastic modeling may also be conducted to resolve
uncertainties.
As described above, the development and application of a computational model
is a long process consisting of many steps. The accuracy and reliability of the end
results rely on manipulations at every step. The developer should approximate the
physical processes reasonably via the mathematical model, derive and code the numer-
ical discretization and solution methods correctly, and verify and validate the model
thoroughly. The user should prepare the data carefully, estimate model parameters
correctly, necessarily calibrate the model, reasonably interpret results, and consider
possible uncertainties.
1.4 CLASSIFICATION OF FLOW AND SEDIMENT
TRANSPORT MODELS
Flow and sediment transport models can be classified in various ways, as described
below.
According to their dimensionality , flow and sediment transport models are classified
as 1-D, vertical 2-D, horizontal 2-D, and 3-D. Flow and sediment transport in natural
rivers are usually 3-D phenomena, which should be more realistically simulated using
3-D models. However, 3-D models are more time-consuming. Therefore, 1-D and
2-D models have been established via simplifications, such as section-, depth-, and
width-averaging, to achieve feasible solutions in engineering practices. 1-D models
study the longitudinal profiles of the cross-section-averaged properties of flow and
sediment transport in rivers. The vertical 2-D models, which may be idealized or
width-averaged, study the (width-averaged) properties of flow and sediment transport
 
Search WWH ::




Custom Search