In this second edition a new chapter has been added covering the
buffeting theory in a finite element format. The motivation for this has
been that a finite element format is becoming more and more dominant in
all areas of structural mechanics. It is streamlined for computer
programming, and it facilitates the use of general purpose routines that
are applicable in several types of structural engineering problems. In
this book the finite element formulation of the problem of dynamic
response calculations follows the general principle of virtual work, a
general principle which may be found in many other text books. While the
buffeting wind load itself has with no trouble been included in a finite
element format, the main challenge has been to obtain a consistent
formulation that includes all the relevant motion induced forces. This
has been important, because, while many structures (e.g. long-span
suspension bridges) may suffer greatly and become unstable at high wind
velocities, the same structures may also benefit from these effects at
the design wind velocity. It is well known that motion induced forces
will change the stiffness and damping properties of the combined
structure and flow system. If calculations are performed for a suitably
close set of increasing mean wind velocities and the changing mechanical
properties (stiffness and damping) are updated from one velocity to the
next, then the response of the system may be followed up to wind
velocities close to the stability limit, i.e. up to response values that
are perceived as unduly large. Finite element calculations may be
performed in time domain, in frequency domain or converted into a modal
format. All these options have been included. Pursuing a time domain
solution strategy requires the use of the so-called indicial functions.
The theory behind such a formulation is also covered, and the
determination of these functions from aerodynamic derivatives has been
included in a separate appendix.
