WIND INDUCED
RESONANT CROSS FLOW VIBRATION
By Arne
Kvitrud, Lauritz Bergendalsgate 1a, 4021 Stavanger.
Paper
presented in 1994, but put on Internet 25.9.2002.
The figures
are not presented here.
INTRODUCTION
Several
problems have occurred offshore in Norway caused by wind induced vibrations.
Wind induced cross flow has caused severe vibrations in two flare booms, and
fatigue cracking in four flare booms and four jackets.
The
guidelines on determination of loads and load effects are recommending the use
of the German DIN 4133 standard, for calculation of cross flow wind vibrations.
This
paper will :
a)
describe the vibrations and its characteristics on Statfjord A and Heimdal,
where the largest vibrations occurred. The vibrations cannot be predicted by
vortex shedding theories of individual elements. The vibrations were caused by
vibrations from wake effects and by vibrations of local frames inside the flare
booms.
b)
review the work done on evaluating the cracks on other flare booms and jackets.
The work have been based on DIN 4133. It will include evaluations of vibrations
of both individual members and frames.
c)
summarize the experience gained from measurements and observations of damping
of individual tubular members, end fixity of members and vibrations of
individual members and frames. Further the paper will describe limitations of
the DIN standard.
IN SERVICE VIBRATIONS
a) During the autumn of 1978 the main
tie of the flare boom of Statfjord A experienced severe vibrations. The severe
vibrations occurred during 3 different days. The wind velocity was about 40
knots when the problems occurred . The amplitudes of the horizontal vibrations
was approximately 100 mm.
The
classic theories for calculating when vibrations would occur, were precise, but
the actual vibrations were up to 20 times larger than the theories predicted.
b)
During the winter of 1984-1985 severe vibrations of the flare boom at Heimdal
occurred. The vibrations occurred both as vibrations of individual members and
as vibrations of frames inside the boom.
The
vibrations of the individual members occurred at velocities which were higher
than expected from model testing of individual members. The vibrations were
later successfully reproduced in model tests in Trondheim. The reason for the
odd behaviour is concluded to be wake interference between the elements. The
response was also higher than predicted by available standards.
The
procedures established in existing standards only consider vibrations of
individual members in a structure. The vibrations of the frames at Heimdal also
highlighted the need to do vortex shedding evaluation of frames.
c)
Based on mountain climbing methods for in service inspection of flare booms,
several cracks have been observed in flare booms during the last years.
On
Statfjord B and C ; 45 cracks have been found in the flare boom. Based on the
use of DIN 4133 ; 44 of the cracks have analytically been found to be caused by
a combination of vibrations of individual members and vibrations of frames.
On
Gullfaks B cracks have been found in eight nodes of the flare boom.
Calculations based on DIN 4133, showed that the cracks most probably were
caused by vibrations of individual members. One also found that additional
loading was introduced when the ratio of natural periods of two connected
members was 1:2.
In
UK sector a similar vibration problem have been reported on Murchison. The
flare boom was designed by the same company that designed the flare booms on
Statfjord, and they are also almost identical.
We
have also reviewed all cracks on jacket substructures above sea water level,
reported during the period 1981-92, based on an assumption that they were
caused by wind induced vortex shedding. The conclusion is that at least four
jacket platforms have cracks caused by wind induced fatigue from vortex
shedding.
DESIGN PROCEDURES
The
use of DIN 4133 has shown to be a useful tool in predicting wind induced vortex
resonant vibrations. The standard have a limitation, however, it can only
handle vibrations of individual members. Experience from vortex induced cross
flow vibrations have shown that it is not sufficient only to review vibration
of individual members. In a real structure the structural elements interact
both with respect to loading (as wakes) and in the response (as vibrations of
frames).
The
proposed CEN standard (EUROCODE 1) is based on the DIN standard, but have some
additional useful guidance to designers and will be a step in the right
direction in predicting vortex induced vibrations. Design procedures to handle
arrays of structural elements (as on Statfjord, Gullfaks and Murchison) is
described.
In
addition to the DIN or Eurocode standards a design procedure have to be
followed using:
* low values of the damping of individual steel members
*
realistic values of end fixity of individual members
*
evaluations of possible frame vibrations have to be included