Detectability Of Structural Failures In Offshore Platforms By Ambient Vibration Monitoring

INTRODUCTION Historically, underwater visual examination by divers has been employed to monitor the structural integrity of fixed offshore platforms. Alternative approaches for inspection are being sought to lessen the cost and hazard of diver operations, especially in deep and hostile waters. An approach being applied in the North Sea makes use of ambient vibration monitoring with above-water accelerometers to identify the lower global modes of platform vibration (e. g., Ref. 1). Changes in the natural frequencies and mode shapes are indicators of possible below water structural breakage. The United States Geological Survey (USGS) has sponsored studies at The Aerospace Corporation to examine the feasibility of this approach. 2 As part of the study, the producing Shell Platform SP-62C in the Gulf of Mexico was instrumented at 17 locations between decks and at the boat landing level for detection of its ambient vibrations during normal producing operation of the platform. Some vertical and angular accelerations were measured, in addition to lateral accelerations. The platform is an eight-leg diagonally braced jacket construction consisting of eight ungrouted main piles and eight grouted skirt piles, and stands in a 327-ft (l00-m) water depth. The interpretation of the random vibration data via spectral analysis is already has been reported, 3 and includes a favorable comparison of the modal parameters derived with subsequent results from forced shaking of this platform. 4 The overall conclusion from this field experiment was that the data acquisition and analysis aspects of the ambient monitoring method are practical, with use of readily available hardware and software. The specifics of the modal identification were as follows:A total of 24 modes were detected between 0.65 and 4.5 Hz. Data acquired during action of 15 to 18 ft (4-1/2 to 5-1/2 m) waves equally were useful as data taken in a calm sea with 2 to 3 it (l/2 to 1 m) swells.The fundamental group of modes (broadside, end on, torsion) respond most prominently. The natural frequencies of these modes in the presence of the rough sea were 1 to 2% lower than those detected in a calm sea. For a given sea state, the frequencies were judged detectable to within 0.5- O. 8% and the ratio of primary late ral motions to within 10%.The frequencies of the second torsion, third end on, first vertical, and three modes involving highly coupled lateral/torsional/vertical motions were judged accurate to within 0.5 - L 2% and ratios of primary motions to within 20%. Although not detected because of tape recorder noise interference, the second broadside mode also is believed to be detectable to this degree of accuracy. The frequencies of a number of other modes could be detected, but mode shape information was poorly determined. The frequency of modes above the fundamentals was not noticeably affected by sea state, most likely because the responses were relatively low.Detection of modes was limited to less than 5 Hz, the frequency regime essentially free of discrete machinery noise.