316
Coastal Engineering: Theory and Practice
Fig. 10.1 Varions steps in obtaining the numerical solution.
Experimental investigations of any physical problem, on the other hand,
are expensive and are often impracticable to conduct on a full scale. Even
though the physical restraints are simplified with the usage of a scaled
model in the experimental investigation, there is always a chance of measurement errors as the extrapolation of the results may be erroneous due
to incorrect approximation of the associated physics and thus similarity
may not be maintained in the results. Numerical solutions are available for
virtually ail problems, although available only at discrète points rather than
a continuous solution in the case of analytical modelling. Once a model is
set up, numerical modelling allows us to simulate different trials with lesser
time and cost than in the case of experiments which involve a separate
set up for each case. Also, it provides complété information whereas in
experiments it is difficult to measure ail the parameters. These models can
handle any degree of complexity, namely hazardous conditions, but the
solution obtained is only as good as the input given.
Since the numerical models are purely based on the approximation of
some mathematical équations, the solution may not be in accordance with
reality. The models cannot guarantee a correct description of the concerned
physical processes involved in the problem if not properly calibrated. Also,
Coastal Engineering: Theory and Practice
Fig. 10.1 Varions steps in obtaining the numerical solution.
Experimental investigations of any physical problem, on the other hand,
are expensive and are often impracticable to conduct on a full scale. Even
though the physical restraints are simplified with the usage of a scaled
model in the experimental investigation, there is always a chance of measurement errors as the extrapolation of the results may be erroneous due
to incorrect approximation of the associated physics and thus similarity
may not be maintained in the results. Numerical solutions are available for
virtually ail problems, although available only at discrète points rather than
a continuous solution in the case of analytical modelling. Once a model is
set up, numerical modelling allows us to simulate different trials with lesser
time and cost than in the case of experiments which involve a separate
set up for each case. Also, it provides complété information whereas in
experiments it is difficult to measure ail the parameters. These models can
handle any degree of complexity, namely hazardous conditions, but the
solution obtained is only as good as the input given.
Since the numerical models are purely based on the approximation of
some mathematical équations, the solution may not be in accordance with
reality. The models cannot guarantee a correct description of the concerned
physical processes involved in the problem if not properly calibrated. Also,