George Forristall - Academia.edu (original) (raw)
Papers by George Forristall
Journal of waterway, port, coastal, and ocean engineering, 1998
Journal of Physical Oceanography, Nov 1, 1987
ABSTRACT The strongest winds, waves, and currents do not necessarily occur at the same time in a ... more ABSTRACT The strongest winds, waves, and currents do not necessarily occur at the same time in a storm, and it is unnecessarily conservative to set design criteria as if they did. We hindcast the most important hurricanes to affect the northern Gulf of Mexico in the twentieth century using the ODGP-2 model for winds and waves and the Mellor-Durbin turbulence closure model for currents. To account for the effect of the joint occurrence of wind, wave and current, we used generic structure models of a rigid platform, a compliant tower, and a tension leg platform to integrate forces. Response functions for the structures were developed through numerical simulations in random waves. Multivariate regression analysis then produced simple polynomial equations for the responses which could be applied to the entire database of oceanographic hindcasts. For the fixed platform, the 100 year base shear found from using this method was approximately 15% less than that which resulted from applying the 100 year oceanographic conditions for winds, waves and currents simultaneously. For a given force level, the hindcasts indicate that a larger proportion of the total force is due to the current than has been thought previously. For the TLP, combined oceanographic criteria which include directional effects were developed. The most likely sea state to cause the 100 year minimum tendon tension was found to occur near the eye of a hurricane. 1 PURPOSE AND SCOPE The environmental forces which a structure must resist cannot be predicted deterministically, so the design for them is necessarily probabilistic. In the face of this uncertainty, a rational determination of the design force level can be made by balancing the probability of failure during a project's planned life against the cost of strengthening the structure. The environmental parameters which would cause the design force then become the oceanographic criteria for new designs. Forces on fixed structures in moderate water depths are largely due to waves, and reliable statistics for waves in the Gulf of Mexico have been available at least since the work of Ward et al (1978). The novel concepts which are being considered for production structures in deep water attract substantial forces from winds and currents in addition to waves. Recent measurements (Price et al, 1991) have shown that hurricane generated currents in deep water can be as large as 7 ft/sec. Deep water oceanographic criteria should thus explicitly include wind and current as well as waves. The design force for many platform types is dominated by waves and a 50 or 100 year return interval wave height is a common design level condition. By extension, one could combine the 100 year wave with the 100 year wind and current. This approach is, however, unduly conservative since the largest wind, wave and current values in 100 years are not expected to occur at precisely the same time. Joint statistics of wind, wave and currents can be developed through hindcasting, but the statistics cannot be used to develop criteria without knowing the relative importance of the various parameters.
Journal of offshore mechanics and Arctic engineering, Jul 7, 2005
This thesis is submitted in conformity with the requirements for the degree of Master of Science ... more This thesis is submitted in conformity with the requirements for the degree of Master of Science (ir.
Journal of ocean engineering and marine energy, Jun 27, 2017
Standard analytical descriptions of individual wave height and period distributions do not agree ... more Standard analytical descriptions of individual wave height and period distributions do not agree well with observations. We show that simple linear simulations from power spectra produce distributions that agree very well with measurements. This is true for measurements in hurricanes, North Sea extratropical storms and bimodal spectra measured in relatively shallow water. Simulations are computationally efficient and useful for applications. The simulations were used to produce parametric wave height and period distributions for Jonswap spectra with varying peak enhancement factors. The parametric distributions also work well for bimodal spectra as long as the spectral peaks are not separated too far in frequency. The wave period of the highest waves in a sea state predicted by the simulations agrees with the period found from the autocorrelation of the spectrum and is slightly longer than the recommendation from the American Petroleum Institute.
Journal of offshore mechanics and Arctic engineering, Sep 13, 2001
Waves in the ocean are nonlinear, random, and directionally spread, but engineering calculations ... more Waves in the ocean are nonlinear, random, and directionally spread, but engineering calculations are almost always made using waves that are either linear and random or nonlinear and regular. Until recently, methods for more accurate computations simply did not exist. Increased computer speeds and continued theoretical developments have now led to tools which can produce much more realistic waves for engineering applications. The purpose of this paper is to review some of these developments. The simplest nonlinearities are the second-order bound waves caused by the pairwise interaction of linear components of the wave spectrum. It is fairly easy to simulate the second-order surface resulting from those interactions, a fact which has recently been exploited to estimate the probability distribution of wave crest heights. Once the evolution of the surface is known, the kinematics of the subsurface flow can be evaluated reasonably easily from Laplace’s equation. Much of the bound wave structure can also be captured by using the Creamer transformation, a definite integral over the spatial domain which modifies the structure of the wave field at one instant in time. In some ways, the accuracy of the Creamer transformation is higher than second order. Finally, many groups have developed numerical wave tanks which can solve the nonlinear wave equations to arbitrary accuracy. The computational cost of these solutions is still rather high, but they can directly calculate potential forces on large structures as well as providing test cases for the less accurate, but more efficient, methods.
Hindcasting has become a generally accepted method of producing long time series of environmental... more Hindcasting has become a generally accepted method of producing long time series of environmental events from which environmental design criteria can be developed. The need for reliable hindcasts has stimulated research into both descriptive and dynamic models of severe storms and the response of the ocean to those storms. Hindcast models can also usefully describe features of the mesoscale ocean circulation such as the Loop Current and its eddies in the Gulf of Mexico. In this review, we examine the hindcast philosophy with illustrations of models for winds, waves, and currents. Areas needing further research are highlighted. These areas include low frequency wind spectra, frictional dissipation of waves in shallow water, hurricane driven currents over the shelf break, and extremes in the mesoscale ocean circulation. The response of a structure generally depends on some combination of environmental criteria. Joint distributions of the oceanographic parameters can be developed from hindcasts, but it is difficult to determine the relative importance of the parameters in generating the response. The most direct way to develop combined criteria is through integration of their effects on sample structures. A number of groups have now developed implementations of this concept, which fortunately seem to give similar results.
All Days, May 5, 1980
ABSTRACT From April 1978 through February 1979, wind and wave measurements were made from the Wes... more ABSTRACT From April 1978 through February 1979, wind and wave measurements were made from the Western Pacesetter II while it was drilling in the Baltimore Canyon area approximately 100 miles off the coast of New Jersey. An anemometer was mounted at the top of the drilling derrick and a Wave rider buoy was moored nearby. Approximately 20 minutes of data each three hours were recorded by a microprocessor based digital recording system onto cassette tapes. The data has been processed to yield average wind speed and direction, significant wave heights, and wave spectra. Seasonal and summary statistics have been developed which show, for instance, that a significant wave height of ten feet was exceeded 8.5 percent of the year. The highest measured significant wave height during the year was 19.5 feet. These statistics should be useful in planning future operations in the area. Study of the high frequency portions of the spectra revealed that over a substantial range, it was inversely proportional to the fourth power of the frequency and directly proportional to the wind friction velocity. The classical inverse fifth power equilibrium range appeared only for frequencies substantially above that of the spectral peak. Using these observations, it was possible to produce a consistent representation of the high frequency tail of the spectrum which should be useful for developing criteria for structural fatigue studies. INTRODUCTION During 1978 and the early part of 1979, Shell was one of a number of operators who tested leases which had been acquired in the Baltimore Canyon area off the coast of New Jersey (Figure 1). The drilling was done from the Western Pacesetter II. Since the entire east coast of the United States is essentially a frontier area for offshore drilling, additional environmental data on the area is very useful for planning future activities. Thus, in addition to participating in the joint industry instrumentation project described by Rose et al. (1979), we decided to make wind and wave measurements from the Pacesetter II. Since most of the instrumentation needed was already on hand, it was possible to start making measurements in April, shortly after the rig moved on site. Because of the simplicity of the instrumentation, a reasonably complete data set was collected even though the system was operated without regular service checks. The measurements add significantly to the marine climatology data base for the area, as well as providing some interesting information on the characteristics of the high frequency tail of the wave spectrum. We begin by discussing the instrumentation and its installation on the Pacesetter II. Our operating experience with this system may help others plan similar measurement programs. The data was first processed to yield histories of the significant wave height and wind speed and direction. From these histories, we calculated seasonal and summary exceedance statistics.
Ocean waves are dispersive and directionally spread, changing size and shape as they propagate. T... more Ocean waves are dispersive and directionally spread, changing size and shape as they propagate. Therefore the maximum crest height over an area in a given length of time will be larger than the maximum crest at a single point. Extreme crest heights are usually calculated from single point statistics, but the designer of a platform is really interested in the probability of a wave crest reaching any part of the deck area. Statistics for the maximum crest over an area have been developed using a combination of analytic theory and numerical simulations. The resulting crest heights are significantly higher than given by point statistics even for relatively small areas. On the other hand, only a small fraction of the deck may be inundated. That fraction can be estimated by a applying a two dimensional version of the NewWave method that finds the most probable shape of an extreme wave.
Journal of Physical Oceanography, Sep 1, 1980
GZ Forristall, EH Doyle, W. Silva, M. Yoshi. In Modeling Marine Systems, vol. II (1990). wave-mud... more GZ Forristall, EH Doyle, W. Silva, M. Yoshi. In Modeling Marine Systems, vol. II (1990). wave-mud-dissipation.
Journal of Geophysical Research, Mar 20, 1975
Journal of Geophysical Research, Nov 10, 1972
Journal of Physical Oceanography, Sep 1, 1978
Offshore Technology Conference, 1979
Wave and heave measurements were made from the Sedco 706 drilling in the Gulf of Alaska in the wi... more Wave and heave measurements were made from the Sedco 706 drilling in the Gulf of Alaska in the winter of 1976-77 and from the Odeco Ocean Prospector drilling offshore of Southern California in the first half of 1977. Wavestaff measurements from the moving semisubmersibles were corrected by adding the heave measured by accelerometers. This wave staff compensation was checked by spectral comparison with data from a Waverider buoy moored off of the Sedco 706 and was found to give excellent results. In addition to increasing the climatological data base for offshore California and Alaska, the measurements were useful for verification of rig motion model tests and computer programs. Prototype data during storm seas has rarely been available for this type of comparison, which is increasingly important since similar methods are being used in the design of tension leg structures. The simultaneous wave profile and heave measurements were analyzed by cross-spectral methods to give the amplitude and phase of the transfer function and the coherence between the signals. The coherences were very high in the energetic part of the spectrum, demonstrating that the heave may be adequately described by a linear transfer function. The measured transfer functions were correctly predicted by the model tests and computer programs.
Storm-driven currents can be an important p a r t of the design hydrodynamic flow field for fixed... more Storm-driven currents can be an important p a r t of the design hydrodynamic flow field for fixed platforms. In addition, the currents which e x i s t during p'latform construction can significantly a f f e c t t h e instalZtxrLon. Planning for t h e i n s t a l l a t i o n of a large platform off the Mississippi Delta was f a c i l i t a t e d by climatologica 2 current data collected with electromagnetic and.Aanderaa meters, which were supported from a semisubmersible d r i l l i n g r i g and from a subsurface mooring. In addition,. Cyc Zesondes and electromagnetic current meters were used t o provzde real time current data during the instaZZation of the base section of the structure. Once a platform is i n place, it provides an e x c e l l e n t s i t e for the s t u d y of near-surface, storm-driven currents and waves. The f a s t response time of e'lectromagnetic current meters makes them seem ideally suited for this application, and their effectiveness has been demonstrated through five years of experience a t t h r e e s i t e s i n t h e GuZf of Mexico. Early problems with reliabi'lity of the meters during long-term,unattended operation have now been most'ly eZiminated. During tropical storm Delia, surface currents over 2 m/sec were measured. The electromagnetic current meters also provide information on the kinematics of storm waves. Comparison o f t h e measured particle velocities with wave theories shows that the directional spread of the wave energy i s important. The measured particle velocity spectra agree with the predictions of 'linear theory t o within a f e u percent over the energetic frequency range, increasing confidence i n the current measurements.
Offshore Technology Conference, 1978
... transport of water due to the Stokes drift can be seen only in Lagrangian measurements ... of... more ... transport of water due to the Stokes drift can be seen only in Lagrangian measurements ... of 24.4 feet and a period of 8.0 secondg determined and compared with the observations as ... Theestimation of the parameters is made by :omparlng the cross spectra actually measured with ...
Journal of waterway, port, coastal, and ocean engineering, 1998
Journal of Physical Oceanography, Nov 1, 1987
ABSTRACT The strongest winds, waves, and currents do not necessarily occur at the same time in a ... more ABSTRACT The strongest winds, waves, and currents do not necessarily occur at the same time in a storm, and it is unnecessarily conservative to set design criteria as if they did. We hindcast the most important hurricanes to affect the northern Gulf of Mexico in the twentieth century using the ODGP-2 model for winds and waves and the Mellor-Durbin turbulence closure model for currents. To account for the effect of the joint occurrence of wind, wave and current, we used generic structure models of a rigid platform, a compliant tower, and a tension leg platform to integrate forces. Response functions for the structures were developed through numerical simulations in random waves. Multivariate regression analysis then produced simple polynomial equations for the responses which could be applied to the entire database of oceanographic hindcasts. For the fixed platform, the 100 year base shear found from using this method was approximately 15% less than that which resulted from applying the 100 year oceanographic conditions for winds, waves and currents simultaneously. For a given force level, the hindcasts indicate that a larger proportion of the total force is due to the current than has been thought previously. For the TLP, combined oceanographic criteria which include directional effects were developed. The most likely sea state to cause the 100 year minimum tendon tension was found to occur near the eye of a hurricane. 1 PURPOSE AND SCOPE The environmental forces which a structure must resist cannot be predicted deterministically, so the design for them is necessarily probabilistic. In the face of this uncertainty, a rational determination of the design force level can be made by balancing the probability of failure during a project's planned life against the cost of strengthening the structure. The environmental parameters which would cause the design force then become the oceanographic criteria for new designs. Forces on fixed structures in moderate water depths are largely due to waves, and reliable statistics for waves in the Gulf of Mexico have been available at least since the work of Ward et al (1978). The novel concepts which are being considered for production structures in deep water attract substantial forces from winds and currents in addition to waves. Recent measurements (Price et al, 1991) have shown that hurricane generated currents in deep water can be as large as 7 ft/sec. Deep water oceanographic criteria should thus explicitly include wind and current as well as waves. The design force for many platform types is dominated by waves and a 50 or 100 year return interval wave height is a common design level condition. By extension, one could combine the 100 year wave with the 100 year wind and current. This approach is, however, unduly conservative since the largest wind, wave and current values in 100 years are not expected to occur at precisely the same time. Joint statistics of wind, wave and currents can be developed through hindcasting, but the statistics cannot be used to develop criteria without knowing the relative importance of the various parameters.
Journal of offshore mechanics and Arctic engineering, Jul 7, 2005
This thesis is submitted in conformity with the requirements for the degree of Master of Science ... more This thesis is submitted in conformity with the requirements for the degree of Master of Science (ir.
Journal of ocean engineering and marine energy, Jun 27, 2017
Standard analytical descriptions of individual wave height and period distributions do not agree ... more Standard analytical descriptions of individual wave height and period distributions do not agree well with observations. We show that simple linear simulations from power spectra produce distributions that agree very well with measurements. This is true for measurements in hurricanes, North Sea extratropical storms and bimodal spectra measured in relatively shallow water. Simulations are computationally efficient and useful for applications. The simulations were used to produce parametric wave height and period distributions for Jonswap spectra with varying peak enhancement factors. The parametric distributions also work well for bimodal spectra as long as the spectral peaks are not separated too far in frequency. The wave period of the highest waves in a sea state predicted by the simulations agrees with the period found from the autocorrelation of the spectrum and is slightly longer than the recommendation from the American Petroleum Institute.
Journal of offshore mechanics and Arctic engineering, Sep 13, 2001
Waves in the ocean are nonlinear, random, and directionally spread, but engineering calculations ... more Waves in the ocean are nonlinear, random, and directionally spread, but engineering calculations are almost always made using waves that are either linear and random or nonlinear and regular. Until recently, methods for more accurate computations simply did not exist. Increased computer speeds and continued theoretical developments have now led to tools which can produce much more realistic waves for engineering applications. The purpose of this paper is to review some of these developments. The simplest nonlinearities are the second-order bound waves caused by the pairwise interaction of linear components of the wave spectrum. It is fairly easy to simulate the second-order surface resulting from those interactions, a fact which has recently been exploited to estimate the probability distribution of wave crest heights. Once the evolution of the surface is known, the kinematics of the subsurface flow can be evaluated reasonably easily from Laplace’s equation. Much of the bound wave structure can also be captured by using the Creamer transformation, a definite integral over the spatial domain which modifies the structure of the wave field at one instant in time. In some ways, the accuracy of the Creamer transformation is higher than second order. Finally, many groups have developed numerical wave tanks which can solve the nonlinear wave equations to arbitrary accuracy. The computational cost of these solutions is still rather high, but they can directly calculate potential forces on large structures as well as providing test cases for the less accurate, but more efficient, methods.
Hindcasting has become a generally accepted method of producing long time series of environmental... more Hindcasting has become a generally accepted method of producing long time series of environmental events from which environmental design criteria can be developed. The need for reliable hindcasts has stimulated research into both descriptive and dynamic models of severe storms and the response of the ocean to those storms. Hindcast models can also usefully describe features of the mesoscale ocean circulation such as the Loop Current and its eddies in the Gulf of Mexico. In this review, we examine the hindcast philosophy with illustrations of models for winds, waves, and currents. Areas needing further research are highlighted. These areas include low frequency wind spectra, frictional dissipation of waves in shallow water, hurricane driven currents over the shelf break, and extremes in the mesoscale ocean circulation. The response of a structure generally depends on some combination of environmental criteria. Joint distributions of the oceanographic parameters can be developed from hindcasts, but it is difficult to determine the relative importance of the parameters in generating the response. The most direct way to develop combined criteria is through integration of their effects on sample structures. A number of groups have now developed implementations of this concept, which fortunately seem to give similar results.
All Days, May 5, 1980
ABSTRACT From April 1978 through February 1979, wind and wave measurements were made from the Wes... more ABSTRACT From April 1978 through February 1979, wind and wave measurements were made from the Western Pacesetter II while it was drilling in the Baltimore Canyon area approximately 100 miles off the coast of New Jersey. An anemometer was mounted at the top of the drilling derrick and a Wave rider buoy was moored nearby. Approximately 20 minutes of data each three hours were recorded by a microprocessor based digital recording system onto cassette tapes. The data has been processed to yield average wind speed and direction, significant wave heights, and wave spectra. Seasonal and summary statistics have been developed which show, for instance, that a significant wave height of ten feet was exceeded 8.5 percent of the year. The highest measured significant wave height during the year was 19.5 feet. These statistics should be useful in planning future operations in the area. Study of the high frequency portions of the spectra revealed that over a substantial range, it was inversely proportional to the fourth power of the frequency and directly proportional to the wind friction velocity. The classical inverse fifth power equilibrium range appeared only for frequencies substantially above that of the spectral peak. Using these observations, it was possible to produce a consistent representation of the high frequency tail of the spectrum which should be useful for developing criteria for structural fatigue studies. INTRODUCTION During 1978 and the early part of 1979, Shell was one of a number of operators who tested leases which had been acquired in the Baltimore Canyon area off the coast of New Jersey (Figure 1). The drilling was done from the Western Pacesetter II. Since the entire east coast of the United States is essentially a frontier area for offshore drilling, additional environmental data on the area is very useful for planning future activities. Thus, in addition to participating in the joint industry instrumentation project described by Rose et al. (1979), we decided to make wind and wave measurements from the Pacesetter II. Since most of the instrumentation needed was already on hand, it was possible to start making measurements in April, shortly after the rig moved on site. Because of the simplicity of the instrumentation, a reasonably complete data set was collected even though the system was operated without regular service checks. The measurements add significantly to the marine climatology data base for the area, as well as providing some interesting information on the characteristics of the high frequency tail of the wave spectrum. We begin by discussing the instrumentation and its installation on the Pacesetter II. Our operating experience with this system may help others plan similar measurement programs. The data was first processed to yield histories of the significant wave height and wind speed and direction. From these histories, we calculated seasonal and summary exceedance statistics.
Ocean waves are dispersive and directionally spread, changing size and shape as they propagate. T... more Ocean waves are dispersive and directionally spread, changing size and shape as they propagate. Therefore the maximum crest height over an area in a given length of time will be larger than the maximum crest at a single point. Extreme crest heights are usually calculated from single point statistics, but the designer of a platform is really interested in the probability of a wave crest reaching any part of the deck area. Statistics for the maximum crest over an area have been developed using a combination of analytic theory and numerical simulations. The resulting crest heights are significantly higher than given by point statistics even for relatively small areas. On the other hand, only a small fraction of the deck may be inundated. That fraction can be estimated by a applying a two dimensional version of the NewWave method that finds the most probable shape of an extreme wave.
Journal of Physical Oceanography, Sep 1, 1980
GZ Forristall, EH Doyle, W. Silva, M. Yoshi. In Modeling Marine Systems, vol. II (1990). wave-mud... more GZ Forristall, EH Doyle, W. Silva, M. Yoshi. In Modeling Marine Systems, vol. II (1990). wave-mud-dissipation.
Journal of Geophysical Research, Mar 20, 1975
Journal of Geophysical Research, Nov 10, 1972
Journal of Physical Oceanography, Sep 1, 1978
Offshore Technology Conference, 1979
Wave and heave measurements were made from the Sedco 706 drilling in the Gulf of Alaska in the wi... more Wave and heave measurements were made from the Sedco 706 drilling in the Gulf of Alaska in the winter of 1976-77 and from the Odeco Ocean Prospector drilling offshore of Southern California in the first half of 1977. Wavestaff measurements from the moving semisubmersibles were corrected by adding the heave measured by accelerometers. This wave staff compensation was checked by spectral comparison with data from a Waverider buoy moored off of the Sedco 706 and was found to give excellent results. In addition to increasing the climatological data base for offshore California and Alaska, the measurements were useful for verification of rig motion model tests and computer programs. Prototype data during storm seas has rarely been available for this type of comparison, which is increasingly important since similar methods are being used in the design of tension leg structures. The simultaneous wave profile and heave measurements were analyzed by cross-spectral methods to give the amplitude and phase of the transfer function and the coherence between the signals. The coherences were very high in the energetic part of the spectrum, demonstrating that the heave may be adequately described by a linear transfer function. The measured transfer functions were correctly predicted by the model tests and computer programs.
Storm-driven currents can be an important p a r t of the design hydrodynamic flow field for fixed... more Storm-driven currents can be an important p a r t of the design hydrodynamic flow field for fixed platforms. In addition, the currents which e x i s t during p'latform construction can significantly a f f e c t t h e instalZtxrLon. Planning for t h e i n s t a l l a t i o n of a large platform off the Mississippi Delta was f a c i l i t a t e d by climatologica 2 current data collected with electromagnetic and.Aanderaa meters, which were supported from a semisubmersible d r i l l i n g r i g and from a subsurface mooring. In addition,. Cyc Zesondes and electromagnetic current meters were used t o provzde real time current data during the instaZZation of the base section of the structure. Once a platform is i n place, it provides an e x c e l l e n t s i t e for the s t u d y of near-surface, storm-driven currents and waves. The f a s t response time of e'lectromagnetic current meters makes them seem ideally suited for this application, and their effectiveness has been demonstrated through five years of experience a t t h r e e s i t e s i n t h e GuZf of Mexico. Early problems with reliabi'lity of the meters during long-term,unattended operation have now been most'ly eZiminated. During tropical storm Delia, surface currents over 2 m/sec were measured. The electromagnetic current meters also provide information on the kinematics of storm waves. Comparison o f t h e measured particle velocities with wave theories shows that the directional spread of the wave energy i s important. The measured particle velocity spectra agree with the predictions of 'linear theory t o within a f e u percent over the energetic frequency range, increasing confidence i n the current measurements.
Offshore Technology Conference, 1978
... transport of water due to the Stokes drift can be seen only in Lagrangian measurements ... of... more ... transport of water due to the Stokes drift can be seen only in Lagrangian measurements ... of 24.4 feet and a period of 8.0 secondg determined and compared with the observations as ... Theestimation of the parameters is made by :omparlng the cross spectra actually measured with ...