Release notes for October, 1999 SeaSoft release The usual caveats about new releases apply: you would be wise to set up a new SeaSoft directory containing this release and save your existing working directories for several months until the new release has had wide exposure and any unanticipated problems that might arise have been resolved. You should copy (not move) any necessary data files into the new release directory so you can retain copies of all data files used in the previous release in their "virgin" state. Don't forget those LOWDAT, USERRAOS and DRFTCOFS files... =========================================================================== >>> Release Highlights =========================================================================== - New version numbers assigned to all software; we now use a single version number that is the same for all applications: Currently SeaSoft version: 4.07. In the future, therefore, it will be possible for an application's version number to change without any actual change in the underlying code for that application. - The FAQ has been significantly expanded. Check it out. It remains very much a work in progress; I have not yet organized or cross-referenced it by simulation or topic, so you will need to search the entire file, using any suitable editor, to find keywords of interest to you. - All manuals have been updated and prepared in Adobe PDF format to facilitate cross-platform viewing and printing - A list of major enhancements, modifications and bug fixes follows: =========================================================================== >>> Implement "Ensemble-wide" statistical measures for peak load/motions =========================================================================== The "storm extreme values" normally quoted in offshore mooring system design analyses (and in the SeaSoft simulations) are, formally, the "most probable extreme" values for the specified environment and storm duration. That is, if one were to analyze an infinite ensemble of "identical" storms (i.e., with equivalent mean environments as characterized by significant wave height and direction, wind speed and direction, etc.), and collect the observed extreme values (i.e., one value per storm) of any dynamical variable (low- or wave-frequency vessel motions, mooring loads, deck accelerations, etc.), one could construct a probability density for the likelihood of observing any particular range of values for that variable. In particular, the value associated with the maximum of this probability density function (the "modal" value) is, by definition, the "most probable peak" value for the specified environment. It is the most likely value to obtain if one were to choose at random a *single* storm in the infinite ensemble. The probability density is not, however, infinitely narrow so that the ensemble-wide extremes of any variable have a mean, median and standard deviation in addition to the modal value. In fact, because of the shape of the ensemble-wide extreme value density, the ensemble-wide "mean" extreme value of any variable is always greater than the "most probable extreme" value. The standard deviation associated with these extreme values is rather sensitive to the number of cycles in the underlying process. Thus, for example, the standard deviation of the extreme low-frequency surge offset across the ensemble may be quite large since there are very few cycles of this process over a typical storm duration. We have begun to output, as a user-specifiable flag option, a semi-quantitative set of extreme distribution statistics including, in addition to the most probable extreme (which has always been an ingredient of the SeaSoft output stream) the ensemble "mean extreme" and the ensemble "most probable plus one standard deviation extreme". Look at LOWOUT and the last page of RANOUT for these. =========================================================================== >>> Implementation of a mooring stiffness matrix option in Catsim =========================================================================== We have recently had repeated requests for a mooring stiffness implementation in Catsim. This turned out to be a vastly more complex undertaking than I had ever imagined with many theoretical, numerical and implementation issues that I won't get into here. There is, however, some related discussion in the FAQ. The present implementation should be considered "beta", although I have spent a good deal of time with it and am not aware of any problems. Some of the implementation difficulties involved numerical precision issues which limited the accuracy of the required finite-difference differentiations. These differentiation calculations *should* be carried out in double precision but are not at present due to compatibility requirements with interacting code modules. We may revisit this problem in the future to upgrade to double-precision calculations, but for now the results are noticeably "noisy" due to this precision problem. =========================================================================== >>> Reworked & enhanced vessel speed option in all simulations =========================================================================== SeaSoft's historical relative-motion capability was geared primarily towards towing simulations; it therefore permitted specification only of a forward speed. This capability has been generalized to permit arbitrary directions of vessel motion relative to the heading of the vessel. The SeaSoft FAQ has a discussion of the use and limitations of this capability. In particular, the new capability can be used to correct built-in vessel RAOs for the Doppler-shifting effects of waves riding atop an underlying current by specifying a vessel speed and direction of motion counter to the imposed current. Discussion from the vessel speed editor help database: Mean motion of the vessel relative to the surrounding fluid affects the frequency versus wavelength relationship of waves as viewed from the vessel frame of reference; that is, for a specified wave length, the encounter period of the vessel with the waves depends on vessel motion. In unusual cases (for example some situations with a large negative forward speed) the encounter period versus wavelength dependency becomes multi-valued (i.e., two different wavelengths can be associated with a single encounter period). This makes the definition of the wave spectrum in irregular wave simulations problematic since the spectrum is defined in a vessel-fixed frame and there is no way of knowing how to apportion the wave energy in a given frequency band between the two associated wavelengths. The vessel speed option should therefore be used with special caution and particular attention given to its affect on irregular wave vessel responses. The vessel speed option permits specification of a direction of motion in the global coordinate system; the usual right-handed coordinate system applies with 0 degrees corresponding to a forward speed condition. Note that vessel motion in the 90 degree direction equates physically to a "current" with a 270 degree "heading". =========================================================================== >>> Reworked treatment of Semisubmersible drift force input requirements =========================================================================== The original built-in wave drift force model for semisubmersibles was implemented in a confusing and unintuitive way by tying the vessel "length" and "beam" directly to the wave drift coefficients. This had several undesirable side effects and has been modified. The user now specifies, in addition to vessel waterline length and breadth, head-on and beam-on "shadowing" factors which contain the subjective degree of shadowing of down-wave columns by their up-wave counterparts. Discussion from the column shadowing editor help database: This item is used in built-in wave-drift force calculations for semisubmersible-type vessels. Briefly, the "shadowing & geometry" factor is a subjective, adjustable constant which is multiplied by the user-specified Beam and Length of the vessel to determine an "effective waterline length" for evaluation of the short wave length limit of the head-on and beam-on wave drift force. This factor should be 1.0 in the limit of *no* inter-column shadowing and flat-faced rectangular columns. For *circular* columns without shadowing, a factor of .67 is applicable, reflecting the favorable geometry of a circular waterline which produces highly non-directional scattering of incident collimated wave energy. To accommodate shadowing, these factors can be further reduced, depending on circumstances, but should rarely be less than about 0.4. Consult the user manuals for additional detail. =========================================================================== >>> Complete overhaul of mooring feedback for buoys in CALMsim & Moorsim =========================================================================== The history of the mooring feedback option (currently only available for buoys) is rather long and gruesome. Until recently, we had not been able to implement a method that worked reliably under all circumstances, the main problem being that any reasonable linearization of mooring system stiffness would too often cause problems in some combination of natural periods and environment. We finally have resolved this issue to our satisfaction by implementing a fully nonlinear energy method (for the wave-frequency mooring feedback) qualitatively similar to what we have previously used successfully to handle problematic low-frequency mooring nonlinearities. The energy method insures that there will never be problems with "runaway" resonant amplification in any degree of freedom of the type that plagued previous attempts. The implementation is complex and we would not be surprised if it undergoes considerable refinement over time; still, it is a massive improvement over earlier attempts and we are eager for "user feedback on the mooring feedback". =========================================================================== >>> Refinements and bug fixes in wave frequency line dynamics model =========================================================================== - Inertial corrections revisit. - Minor bug fixes to wave-frequency loads reported in RANOUT and SNAPOUT. - Large-amplitude corrections revisit. - Improved calculation of small-amplitude catenary spring constants used in the wave-frequency line load model. - Added an option to use fairlead motions along the normal to the surface of constant energy (rather than along the fairlead line tangent) as input to the wave-frequency line load estimator. This is to provide compatibility with early versions of "mooring feedback" simulations for small CALM-type buoys. In most cases, the "energy surface" option will produce differences in the wave-frequency load estimates of 20% or less although in unusual circumstances the differences can be larger. This option should be avoided unless you are quite confident you understand what you wish to accomplish by using it. =========================================================================== >>> Miscellaneous bug fixes & updates: =========================================================================== - Fixed a bug that caused, in some circumstances, "extreme" offset distances to be used for line load estimates instead of "characteristic" offsets in SNAPOUT & RANOUT. - Fixed a bug that produced small errors in the reported RANOUT/SNAPOUT wave-frequency line load variation and total load estimates. - Fixed a error in the MEANOUT report of the global X and Y mooring moments at equilibrium - Fixed yaw damping physical units in LOWOUT. - Fixed spread moor bug that in some circumstances could cause a memory access error in the nonlinear correction routine NLNCOR. - Fixed a bug that caused execution errors for weightless lines in some circumstances. - Fixed a logically inconsistent treatment of current effect and forward speed on natural damping levels in roll, pitch and heave in Semisim. - Fixed numerous bugs in CALMsim that prevented proper program execution when using imported files in some circumstances. - Fixed a bug in CALMsim that produced a slightly erroneous offset-versus-system energy relationship and therefore small errors in the reported characteristic and extreme low-frequency offsets and loads. - Fixed a bug in Catsim that prevented proper reporting of "vessel-based" load components when trim or heel had been applied. - Eliminated an internal precision judgement in Catsim; previously, floating point numbers smaller in magnitude than a certain value were displayed as zero for cosmetic purposes. This caused output such as total vessel moment for a zero offset in a symmetric moor, which should be precisely zero, to be displayed as zero even though the calculated value, although extremely small, is nonzero due to numerical precision limitations. It was felt that an arbitrary cutoff was risky and that the user was in a better position to make this judgement on a case-by-case basis. Displaying these "calculated zeros" also confers a notion of the level of floating point precision actually achieved. - Fixed a bug that caused some non-fatal editor exit warnings to be invisible to the user. - Numerous cosmetic, on-line help, and output descriptor changes in all simulations. - Minor Y2K modifications in date output format. =========================================================================== Ends release notes of October, 1999 ===========================================================================