Role of Dominant Wave tn Spectrum of Wind-Generated Waves 
REPLY TO DISCUSSION 
E.J. Plate 
Untverstty of Karlsruhe 
Karlsruhe, Germany 
I think the two questions amounted to about the same thing. 
I shall try to answer them indirectly by giving my opinion on how 
best to represent the sea state in a mathematical or laboratory mo- 
del. 
I must speak as a coastal engineer because we are usually 
concerned with structures that have very few degrees of freedom. In 
coastal engineering we can usually identify only one or at best a small 
number of degrees of freedom and of corresponding natural frequencies 
or eigen-frequencies of our structures. Therefore I would recommend 
that if one does laboratory studies of the vibrations of a structure one 
should always set the wave conditions for the laboratory in such a way 
that the most critical eigen-frequency of the structure corresponds to 
the peak in the spectrum of the wave, if this is at all possible. Of 
course, if this natural frequency is of the order of 25 Hz, itis point- 
less to try to get 25 Hz waves. But if itis, for example, of the order 
of 0.1 Hz, may be the most important design case arises when the 
frequencies of the dominant wave and the natural frequency of the 
structure match. This is somewhat different from present usage 
among coastal engineers. 
I should like also to say a word against the necessity for using 
so-called random waves for modelling the sea state in a laboratory. 
You have probably all heard that it is becoming more and more fas- 
hionable to use a random wave generator instead of the older sinusoi- 
dal wave generator to model the forces on structure. In my opinion this 
is an ill-considered move for the simple reason that the random wave 
components you are generating are component waves - that is, each 
one of these waves is travelling at its own celerity. Therefore, these 
waves show interference pattern - that is, the small waves modulate 
the big waves, and vice versa, and you get a breaking of waves owing 
isi95 
