At-Sea Test Of The Structural Response Of A Large-Diameter Pipe Attached To A Surface Vessel

L.A. Vega, Gérard C. Nihous

Offshore Technology Conference · 1988 · 24 citations · 0 references

Concepts

Abstract

ABSTRACT In 1983, the U.S. Department of Energy sponsored an at-sea test to evaluate the structural response or a large 8 foot diameter fiberglass-reinforced plastic pipe suspended from a barge. The reduction and analysis of the data collected showed a strong causal relationship between barge horizontal motions and pipe bending strains. In addition, experimental values of pipe drag and added-mass coefficients were provided for Reynolds numbers in the supercritical flow regime, whereas peak responses were determined to be Rayleigh distributed. INTRODUCTION In the late seventies, the analysis of large scale marine risers and pipes had relied on computer-aided studies and laboratory tests. In order to facilitate the design of cost-effective structures, the National Oceanic and Atmospheric Administration implemented the fabrication, transportation, deployment and operation of a heavily instrumented 8-foot diameter, 400foot long, fiberglass-reinforced plastic pipe suspended from a barge by means of a double gimbal, as shown on Figure 1. This at-sea test was performed off Honolulu in the Spring of 1983. Time-history records of wave height, water velocity profile, barge motions, gimbal angles and strains at nine stations along the pipe were obtained under varying operational and environmental conditions. Instruments used for this test were, respectively, a waverider buoy, a current meter, accelerometers and strain gauges. Data quality and data reduction are discussed in details in a previous report by Vega, Nihous and Hoech, 1985. A minimum record length of 550 seconds at a sampling period of 1 second proved sufficient to insure statistical stability. Moreover, the data was estimated to be stationary, from the relative invariance of such parameters as mean and standard deviation over contiguous 550-point records. The normality of the data was determined with a Chisquare goodness-of-fittest, following Bendat and Piersol, 1971. SPECTRAL ANALYSIS OF BARGE MOTIONS, PIPE STRAINS AND BARGE-PIPE COUPLING Power spectral density estimates S (W) were derived from time-history records by the correlation function method (Otnes and Enochson, 1978) for 64 frequencies. The upper limit is the Nyquist value of 3.14 rad/sec corresponding to a period of twice the sampling rate. A theoretical relationship between record length and the interval into which the ratio S (W) / S (W) falls with a chosen probability is provided in Otnes and Enochson, 1978.1 Figure 2 shows the 95% confidence limits for 64-point spectra and data sampled every second. The uncertainty band should be thought of as conservative because it applies pointwise; globally, the constraint holds. Table 1 lists the record names which are referred to in this paper, as well as their length. Figure 3 shows the wave energy spectra for LONG7 and LONG5, which correspond to the most active environmental conditions met during the at-sea test. The pipe bending response, as on Figure 4, is highly resonant with well-defined peaks for the first two vibrational mode shapes.