Page 90 - Dynamic Loading and Design of Structures

Page 68

perturbation of the wind speed at a point about the mean; the hourly mean value V is the ideal
reference for the large storms typical of temperate climates. If a Cartesian co-ordinate system
is referred to the mean wind direction, (x alongwind,y horizontal crosswind, z vertical) a
corresponding component system (u, v, w, respectively) can be applied to the velocity
fluctuation. Notation σ, etc. is used to denote the r.m.s. fluctuation, and the intensity of
u
turbulence I=σ u/V; σ u as well as V is a function of height above ground z. The ‘equilibrium’
values resulting from a very long fetch of uniform roughness are well established together
with reasonable evaluation of the development over changes in roughness (Harris and Deaves,
1981; Cook, 1985).
The UK general code of practice for wind loads, BS6399 Part 2 includes a socalled
‘directional procedure’ with tabulated coefficients relating the local mean speed V(z) and
intensity of turbulence I(z) to the basic storm strength (Vb in the notation of BS6399) as a
function of the local terrain and topography. Factors Sc and St give V and I, respectively, for
locations in open country, as a function of distance from the coast in each selected wind
direction, with the possible addition of allowance for the influence of ground contour
(topography, Sb,). Further corrections (factors Tc and Tt) are given for sites in urban or forest
terrain. BS6399 continues with procedures to assess the correlation of gust action over the
extent of the structure as a static load process and a simple generalized factor for dynamic
augmentation of response.
To proceed further to address the dynamic effects of gusts, in the sense of effects
influenced by the inertia of the structure, the methodology is extended by representation in the
frequency domain using the Fourier integral transform as outlined in Section 3.1.2. The
Fourier integral (spectral) approach can also be used in assessing the spacial correlation of
quasi-static pressures, which is dominated by the action of frequency components well below
any resonant frequency of the structure, where inertial effects are negligible. Some of the
approximations made in the dynamic analysis given in this chapter become questionable at
low frequencies, and the writer advocates instead a direct approach to this part of the
windload problem, either by formal static correlation analysis (Wyatt, 1981) or by the
postulate of a critical gust duration proportional to the quotient of the size of the loaded
surface in question with the mean windspeed (Cook, 1985). This so-called ‘TVL’ approach
(gust averaging time, windspeed, loaded length) is developed in BS6399 Part 2.
The distinctive characteristics of gust response, by contrast especially with earthquake
effects (see Chapter 4), are:
●long duration, the storm persisting near peak intensity for duration of order 1 hour;
●a coexistent mean load, permitting no reversal of inelastic deformations;
●the forces are randomly spatially variable over the structure;
●effects are significant over a frequency range extending down to a few cycles per hour.