RIFLEX Release Notes

A log of all signal to and from the external controller is now available. The file <turbine name>.log is stored in the analysis folder.

The angular position of blade 1 where 0 is with blade 1 pointing upwards is now available to the external wind turbine controller.

Computation of wave kinematics in updated position has been made available for coupled models.

Cross flow and inline displacement envelope curves are stored on mpf-files.

RIFLEX now supports multiple wind turbines with external control systems. Previously only one wind turbine could have an external control system.

The combination of the option iopstr=1 (stresses from curvature), zero cross sectional diameter or elastic modulus is illigal input. That is both cross sectional diameter or elastic modulus must be larger than zero.

Illegal input will now give an error message and program termination.

Hydrodynamic loads may now be exported for specified elements. The added mass contribution is not included. For elements with time domain VIV loading, the relative velocity and the different load contributions may also be exported.

The default local axes method for wind turbine blades is not followed if the blade line is not pre-bent or pre-twisted.

This is corrected.

Long-standing error.

The swell realization in a coupled model has been found to be incorrect, ie. that the specified swell spectrum has not contributed to wave kinematics. The error has been corrected. Longstanding error!

The allocation of some temporary arrays could fail at early in static analysis. This was encountered for analyses with a very large number of elements and caused the analysis to stop without any error message or warning.

The configuration data is now passed in a configuration file to the Bladed wrapper using a configuration file. This allows longer file names to be used.

The original flat bottom formulation has been available in RIFLEX standalone by choosing IBTANG = -9. It is not available in SIMA.

This option has been deprecated since RIFLEX 4.10.0 and will be removed in RIFLEX 4.26.0.

The CARISIMA riser-seafloor contact has been deprecated since RIFLEX 4.16.0 and will be removed in RIFLEX 4.26.0.

Hydrodynamic load models for partially submerged and net cross sections were introduced in RIFLEX 4.22. These load models replace the partially submerged and net cross sections and may be used with both regular and irregular waves.

In SIMA the partially submerged cross sections CRS3 and CRS4 are chosen by selecting the option Use formulation for partly submerged cross section. The corresponding load models are chosen as Morison partly submerged cross section.

Similarly, use of the Fish net cross section (CRS6) may be changed to selecting the Net load model.

The partially submerged and net cross sections are currently deprecated and will be removed in RIFLEX 4.26.0.

The simplified modelling of kill and choke lines attached to a tensioned riser by including them in the riser elements will be removed in a later version of RIFLEX. This functionality is currently available through the variable NAKC in ARBITRARY SYSTEM AR and then giving the input described in Description of kill and choke lines.

Deprecated since RIFLEX 4.10.0.

The hydrostatic stiffness of beam and bar elements crossing the surface is now included in eigenvalue analysis.

Corrections to the net load model that was introduced in 4.22.0.

If the number of kinematics quantities on the file for each kinematic node is not specified, the value will be found from the specified column numbers.

This correction allows wave kinematics to be read from file in SIMA 4.2.

Time domain VIV load model is available for stress joints.

Too large time step gives a Nyquist frequency smaller then the largest frequency in the tabulated spectrum.

The error message is improved.

The time domain VIV load model is available for thin-walled pipe cross sections (CRS0).

The default value for number of slices in core has been increased from 10 to 800. This is expected to increase the CPU efficiency when the IECWind (DTUMann) format is used in the analysis.

A hydrostatic/hydrodynamic load model that accounts for loads acting on a net panel has been made available for axisymmetric cross sections (CRS1). Both lumped and consistent formulations are available, and the load type may be used in dynamic analyses with regular or irregular waves.

The load formulation is based on the Fish net cross section (CRS6) and can now be used with irregular waves.

A hydrostatic/hydrodynamic load model that accounts for partially submerged cross section has been made available for thin-walled pipe cross sections (CRS0), axisymmetric cross sections (CRS1) and double symmetric cross sections (CRS2) as a new hydrodynamic load type. This load model is available for both beam and bar elements and may be used in dynamic analyses with regular or irregular waves. Both lumped and consistent formulations are available.

The new load model will give similar results to the existing partly submerged cross sections (CRS3 and CRS4) which could not be combined with irregular waves.

The irregular wave surface was not visualized when the wave kinematics were computed in updated position. This is now available.

An option have been added to allow the added mass and damping to be given in non-dimensional format.

Allow to perform eigenvalue calculations using cross sections with the MacCamy & Fuchs wave loads.

The external wind turbine control system may now also include Nacelle yaw control. Currently only one wind turbine can have an external control system.

When using wind spectra in coupled analysis, i.e. iwityp 1, 2, 3, 4, 5 and 6, the wind velocity time series is now reported for the hub position. Previously the wind velocity time series were reported for the reference height of the wind coefficients.

Wind velocity for turbine blades will now be computed by taking the blade element vertical position into account. Previously the wind velocity at wind coefficients reference height was used for all blade elements.

This means that results will change slightly compared to older versions.

The new behavior is consistent with how other calculations are made with other wind types, i.e. iwityp 10, 11, 12, 13 and 14.

The new option IOP_ARCLEN may be used to choose whether the arclength is accumulated throughout the whole system or starts at zero for each line. The arclength was previously always accumulated over all lines. This will continue to be the case if this option is not given or if IOP_ARCLEN = 0 is specified.

The arclength is mainly used for results on the MatrixPlot files.

Incorrect pointer caused error in eigen value analysis with multiple bodies. This is correceted.

Longstanding error.

The boxes were earlier placed with the lower left corner where the lower right corner should be. The location in x-direction used the opposite sign of what the user gave.

This is now corrected, and is in line with the reading of wind boxes as described in Turbulence for the IEA Annex 30 OC4 project (Larsen, T. J. "Turbulence for the IEA annex 30 OC4 project." IEA: Paris, France (2011).)

Longstanding error.

The wind field coordinates was not consistent with Turbsim’s definition if the height and width of the grid were different.

This is corrected.

If a wind profile was defined and the wind load group is not activated in the static loading, the static parameter variation failed.

This is corrected.

Error in printing diffracted kinematics points to Dynmod result file is corrected. The incorrect points was listed. The error is related to the print only.

Longstanding error.

If the additional mass for a nodal component attached to a bar element was specified in the global system, the added mass contribution to mass proportional damping (Rayleigh damping) was not included. This has been corrected.

Note that mass proportional damping will dampen rigid body motions and is therefore normally not used.

A possible error in analyses in which some turbines had nacelle yaw control and some did not has been corrected. A wind turbine without nacelle yaw control could turn off yaw control for subsequent wind turbines.

Corrected the scaling of the input yaw threshold for the internal yaw controller. User input is given in (deg^2 s) and is used internally in (rad^2 s). The value was incorrectly scaled with pi/180 instead of with (pi/180)^2.

The internal yaw controller may now have a time step that is larger than the simulation time step.

An option has been added to toggle storage of results for seafloor and soil layer profile contact elements. These results are available from nonlinear simulations only. If storage is specified, they will be stored for all seafloor / soil contact elements attached to beam or bar elements for which force storage is specified.

The results were previously exported if force storage on an additional file was specified; i.e. if IFORFM was nonzero. The results will now be exported if the new option IBOTFM is nonzero.

The default is no storage of seafloor / sol contact results, so the input file must be changed if these results are wanted.

The dynamic input must be modified if wave kinematics are to be read from file. The number of kinematics quantities on the file for each kinematic node must be given. Either 7 or 8 kinematics quantities may be read for each node depending on whether the dynamic pressure is included on the file or not.

Note that if a wave kinematics file with 7 quantities per node is replace by one with 8, the maximum number of columns and the first column number for each node that kinematics are to be read for .must also be changed.

The old, undocumented OUTMOD data group TIME FATIGUE LIFE has been removed. The OUTMOD data group TIME FATIGUE DAMAGE can be used instead.

Deprecated since RIFLEX 4.16.0.

Using the SIMO feature Generic External Control System would previously cause RIFLEX to terminate with an error. This has now been resolved.

Error since RIFLEX 4.20.

Visualization using regular waves in coupled analysis failed to work. This has been solved.

Error since RIFLEX 4.20.

Corrected an error that could occur for multiple wind turbines. If more blade elements were specified for blade storage than the number of blade elements in the last specified wind turbine, the simulation could fail.

Corrected an error in some results for the last, i. e. outermost, element of a blade: - Annulus average axial induced wind in rotor system - Annulus average tangential induced wind in rotor system - Annulus average axial force in rotor system - Annulus average tangential force in rotor system These results are only stored if medium or maximum blade storage is specified.

MacCamy & Fuchs loads can now be used for seastates with both wind sea and swell spectra. Previously, MacCamy Fuchs loads could only be used for a wind sea spectrum.

The line ID is now written to the key-file key_<prefix>_supportforce.txt.

If more than one line is connected to a supernode the resultant force will also be reported. The line ID for the resultant force is given as ALL . The element type for the resultant force is given the text .N.A.

Corrected an error when the lumped load and mass formulation is used with partly submerged and net cross sections, i.e. CRS3, CRS4, CRS5 and CRS6 cross sections.

Error since RIFLEX 4.18.

Corrected an error that could occur for nonlinear time domain simulations with contact elements and sparse matrix storage format. The error caused overwriting of the internal sparse matrix bookkeeping array and was observed to cause program termination.

Error in 4.19 development versions linked after 2021-02-17 and the unreleased internal versions 4.20.0 and 4.20.1.

Coriolis loads and centripetal loads, i.e., load contributions due to rigid motion kinetics, have also been included for the general cross section (CRS7) in combinations with consistent load and mass formulation.

Coriolis loads and centripetal loads, i.e., load contributions due to rigid motion kinetics, have been included for the general cross section (CRS7). These load contributions are only included for lumped load and mass formulation. This functionality has been implemented for analysis of operational wind turbines. These load contributions are normally insignificant for analyses with moderate dynamic rotational speeds, e.g., normal riser analyses.

There is no longer a limit to the number of wind turbines that can be specified. In the current software version, however, only one of them can have an external blade pitch / electrical torque control system. The dynamic load conditions wind events, shutdown and blade pitch faults can only be used for systems with one turbine.

Allows storage of SIMO body and support vessel support forces to file, see the new DYNMOD data group SUPPort FORCe STORage.

Allows modelling of soil embedded structures based on PISA methodology.

Intended for wind turbine monopile foundations with a large diameter-to-embedded-length ratio. Introduces the concept of a soil layer profile and PISA type soils: clay, sand and general Dunkirk sand. Connected lines gain stiffness with respect to lateral displacement and shear deformation.

Nacelle yaw control is implemented. The controller rotates the entire rotor-nacelle assemble (RNA), aligning it with the dominating wind direction.

Coriolis centripetal load has been added for general cross sections CRS7, i.e. the effect of rotational speeds about 2 local element axes that gives rise to moment about the 3rd axis. This functionality was mainly implemented for analysis of operational wind turbines. It is normally insignificant for analysis with moderate dynamic rotational speeds; e.g. normal riser analysis.

The cross sectional damping specified in INPMOD can now differ for axial, torsional and bending dofs in VIVANA.

The curvature-related damping specified in VIVANA is applied to local bending only.

The input to the internal wind turbine controller is now read with the FREAD package which is used for other RIFLEX input.

Visualization of regular wave is enabled.

The visualization is an approximation of the input wave used in the analysis. A single component wave spectrum is used to generate the wave and slight deviation in period may occur.

The initial in-line load model is a cross-flow induced in-line load term. An alternative load model has been implemented where the in-line load term is calculated independently. This also allows simulation with in-line VIV loads only.

A new input format for the time domain VIV load options and coefficients has been added. Old inputs may still be used in 4.20, but the old format will be removed later.

Note that the time domain VIV load model is currently restricted functionality and that a license is required.

Temporary files used during simulations are now stored in the current working directory and not the system’s TEMP location.

DYNMOD will now calculate the necessary size of the file address array for the ifmdyn file. The environmental variable RIFLEX_MAXDYN_IFNDYN is therefore no longer used.

The Bak factor is given the default value of 0.1. The Bak factor will only be used when tower drag effect is included. If the effect of tower drag is not applied, the Bak factor is automatically set to 0. Previously there was no default value for the Bak factor.

General cross section (CRS7) with eccentric mass center applied in the model of blades of an operating wind turbine, has given inaccurate responses. This is because the Coriolis loads and centripetal loads have been missing in the total load formulation. This has been corrected.

Blade pitch fault specified for the first blade in a wind turbine using the internal control system have given incorrect results. This is because the pitch of the first blade was used as feedback to the internal control system. No error for blade pitch faults specified for other blades.

The blade pitch sent into an external control system has been pitch values with the fault applied.

Error since blade pitch fault condition was introduced in RIFLEX 4.12 (SIMA 3.5).

If the added mass of an attached body is given in the global system, it has not contributed to the Rayleigh mass-proportional damping.

Also, the MacCamy & Fuchs added mass from radiation force and the added mass at infinite frequency from potential wave flow theory have not contributed to the Rayleigh mass-proportional damping.

This is corrected and these contributions are now included.

Note that mass-proportional damping should be used with caution as it will give damping for long response periods.

The default procedure for local element axes of vertical elements is now also used for elements pointing downward. Potential numerical problems will be avoided. The local axes and results in the local element system may change for downward-pointing vertical elements.

The description of element local axis in the user manual is updated in wording and with examples and figures.

The reading of fluctuating 3-component wind fields on IECWind format files has been improved. This allows more wind field data to be in memory at the same time without encountering program termination with a stack overflow error. The number of crossectional planes, slices, in memory or the number of points in each plane may therefore be increased.

The tower shadow above the tower top is corrected. The tower shadow no longer gives a jump in the wind velocity when the blade segments are at the same height as the tower top (in negative y-direction). The wind velocity is also closer to free wind velocity when the blade is pointing upwards.

An error caused simulation with user specified stall points to fail if foil coefficients were given for a single Reynolds number. This has been corrected.

Error since RIFLEX 4.18.0.

There was an error if the specified fatigue limit for the SN curve, i. e. the stress level for which the SN curve becomes horizontal, was larger than the SN curve value for a single cycle. In this case the fatigue limit was disregarded and the fatigue calculation was preformed as if no fatigue limit was specified.

The error has been corrected.

Longstanding error.

It is not recommended to use the accelerations exported to the external wind turbine controller directly. Instead, the external controller should itself derive an improved estimate of the accelerations based on the displacement values from several time steps.

The inaccuracies in the calculated accelerations are a result of the Newton-Raphson iteration methods used in the time domain integration.

Accelerations are not used in the internal wind turbine controller.

The default local axes method for wind turbine blades is not followed if the blade line is not pre-bent or pre-twisted. Long-standing error.

Workaround: use local element axes for the wind turbine blades or define a negligible, but non-zero twist or bend for the blade line type.

Data for static restart is now only stored if the data group STORE RESTART DATA is specified on the static input file. Restart data was previously always stored. The _stamod.inp file must therefore be modified for static restart to remain available.

This change is made to reduce file IO and to save space in the working directory.

The input to external wind turbine pitch controller is changed. The specification of additional nodal and element responses to be passed to the external controller, the "additional measurements", is now given on a separate input line. This line comes after the three lines specifying the Java jar file, the Java class name, and the configuration file.

Note that the line with nnod_meas and nel_meas must always be given if an external pitch controller sis used. A blank line is equivalent to specifying both as zero.

The change was mase in the RIFLEX development version 4.19.1.

Removed the default values for - air density (AIRDEN) and - water density (WATDEN) in the ENVIRONMENT CONSTANTS data group in INPMOD - minimum yield stress (SMYS) and - modulus of elasticity (EMOD) in the CODE CHECK CURVES data group in OUTMOD. The user must now specify these values.

The previous default values for AIRDEN and WATDEN were only valid if the units `m' and `kg' are used.

The previous default values for SMYS and EMOD were only valid if the units `m' and `kN' are used.

A new input format for the time domain VIV load options and coefficients has been added. Old inputs may still be used in 4.20, but the old format will be removed later.

Note that the time domain VIV load model is currently restricted functionality and that a license is required.

The original flat bottom formulation will be removed in a later version. It may be specified with IBTANG = -9 in RIFLEX 4.10.x.

Deprecated since RIFLEX 4.10.0.

The old, undocumented OUTMOD data group TIME FATIGUE LIFE will be removed following the 4.16 release.

The OUTMOD data group TIME FATIGUE DAMAGE should be used instead.

Deprecated since RIFLEX 4.16.0.

The simplified modelling of kill and choke lines attached to a tensioned riser by including them in the riser elements will be removed in a later version of RIFLEX. This functionality is currently available through the variable NAKC in ARBITRARY SYSTEM AR and then giving the input described in Description of kill and choke lines.

Deprecated since RIFLEX 4.10.0.

The CARISIMA riser-seafloor contact is deprecated and will be removed in the future.

Deprecated since RIFLEX 4.16.0.

Increase maximum arrays in ifndyn to 100million.

The maximum length of airfoil identifiers was not consistent in code and documentation. The maximum length was 8 characters and not 64 as described in the airfoil library file description. Long identifiers were truncated to 8 characters. This could result in the wrong airfoil being used in the analyses.

The length is now increased and standardized to 32 characters in both SIMO and RIFLEX.

Also, a potential error that could cause data to be overwritten when reading the airfoil library file has been corrected. This error has not been observed.

Incorrect results were obtained for air foils if the air foil system was rotated with regard to the local element system; i.e. if ROTFAX was not zero. The error has been corrected.

Longstanding error.

An error resulted in airfoils with user-specified stall points always operating in dynamic stall. This has been corrected.

Error since RIFLEX 4.6.0

When MacCamy-Fuchs wave loads were calculated for a regular wave given as a tabulated wave spectrum, numerical inaccuracies could lead to the wave frequency being missed in the calculations. This resulted in the MacCamy-Fuchs loads being zero.

Error since RIFLEX 4.18.1, but also present in 4.16.7. The error has been corrected. Some changes in results may occur. The changes will be insignificant if the length of the pre-generated wave time series is reasonably long.

Morison hydrodynamic loads had to be specified to include the effect of eccentric mass center in calculating inertia forces.

It is in general allowed to specify no hydrodynamic forces

The error applied to lumped formulation only.

The potential flow library file name can now start with slash.

Longstanding error.

The performance in pre-generation of MacCamy-Fuchs wave loads has been significantly improved.

The SYROPE model for fibre ropes has been implemented in the new cross section component FIBR.

Reference: Falkenberg et. al.: The SYROPE Method For Stiffness Testing Of Polyester Ropes. In: Proc. of the ASME 2018 37th International Conference on Ocean, Offshore and Arctic Engineering (OMAE). 2018

The requirement to identical blades in a wind turbine model have been relaxed. The wind turbine blades have still to be identical with regarding the element distribution, foil profile description and aerodynamic coefficients along the blades. However, the mass and stiffness distribution may be different.

An option to apply stiffness proportional damping based on material stiffness only has been implemented. This option should be used for wind turbine blades because the geometric stiffness will introduce damping of the rigid body motion. This will also ensure symmetrical behaviour of the blades when constant Rayleigh damping is specified.

A signal named Electrical generator torque is a part of the Wind Turbine output. This is however not on the high speed shaft, but on the low speed shaft.

The name is changed from Electrical generator torque to Mechanical generator torque on LSS

Airfoil coefficients may now be given for a single Reynolds number. They will then be used for all Reynolds numbers.

Extrapolation outside the range of Reynolds values given will now be allowed. The value for the closest Reynolds number will be used; i.e. flat extrapolation.

The import of the airfoil library for foil load coefficients has been improved. This will reduce the analysis time for both static and dynamic analysis. The reduction will be most noticeable in static analysis.

The algorithm for generating pseudo-random numbers may be selected by the user. The "mersenne twister" is the recommended method and should be used unless backwards compatibility with previous versions is required. Note that the default value may change in a future release. The choice of random number generator will apply to: - generation of irregular wave time series - initial phase angles for time domain VIV loads specified for cross-sections in INPMOD - generation of phase angles for application of harmonic loads from a VIVANA frequency domain analysis.

It has been identified that the legacy method can give non-gaussian and non-stationary wave elevation in SIMO for short crested waves with more than about 30-50 discrete directions, depending on wave spectrum and simulation duration. By choosing the mersenne twister, these issues are avoided.

For coupled analysis, wave time series will be generated using the random number generator specified in SIMA.

Stiffness-proportional damping was previously based on only the material stiffness for elements that were - axisymmetric cross section (CRS1) specified with with hysteresis generated by an internal friction moment at reversed curvature - general cross section (CRS7)

To reproduce dynamic results for models containing either of these, the stiffness contribution to damping has to be specified as based on the material stiffness only.

The input to the internal wind turbine controller is now checked more thoroughly. Some input that was previously accepted gave errors in the subsequent analysis. Illegal input will now give an error message and program termination.

Corrected an error that caused MacCamy-Fuchs radiation forces (inertia and dissipation forces due to specified added mass and damping coefficients) to contribute even for elements above the mean water level (dry elements).

Corrected an error that could cause abrupt stop during generation of long second order wave time series.

In rare cases without current, uninitialized values could occur for current velocity at the nodes. These could potentially result in incorrect hydrodynamic loads in dynamic simulations and thus to incorrect dynamic results.

The error was avoided if a current with zero velocity was included in the simulation.

Longstanding error

Corrected an error in coupled analysis when a 6-dof SIMO body is attached to an element system that only contains bar elements. The error could cause overwriting of nodal coordinates and give incorrect results.

No error if at least one element in the system has bending stiffness, for example if a stiff beam element was used to connect the SIMO body to the rest of the element system.

Error since RIFLEX 4.16.0.

Corrected the error that meant that only the first four values of a tabulated wave spectrum were used.

Error since RIFLEX 4.10 (2017-03-21).

If a boundary change was used to change a master node from fixed to free on the first step of static analysis, the slave nodes would end up with the same coordinates as the master node.

Longstanding error.

The damping contribution to the radiation load was not included. This has been fixed. Both the added mass and damping terms now contribute to the radiation loads.

Error since RIFLEX 4.14.0.

The peakedness factor gamma in Torsethaugen spectrum is restricted to be greater than, or equal to, 1.0 since lower values caused an illegal numerical operation and result in NaN in results. The problem has been encountered for seastates with extremely small significant wave height.

Long-standing error

The inner pressure is now applied at end 2 if specified. The pressure can be changed for more than one MRL.

The time domain VIV loads were incorrect if inconsistent mass and force units were used, i.e. GCONS different from 1.0. The error would normally be obvious, e.g. with the units kN and kg, GCONS is 0.001 and the applied loads were too large by a factor of 1000.

The error was avoided when running in SIMA as SIMA always uses a consistent set of units.

Error since RIFLEX 4.14.0.

The in-line term of the time domain VIV loads has been calculated using the relative velocity at the first end of the element at both ends. This has been corrected.

Note that the time domain VIV load model is currently restricted functionality and that a license is required.

Error since RIFLEX 4.14.0.

Corrected an error in coupled analysis when a 6-dof SIMO body is attached to an element system that only contains bar elements. The error could cause overwriting of nodal coordinates and give incorrect results.

No error if at least one element in the system has bending stiffness, for example if a stiff beam element was used to connect the SIMO body to the rest of the element system.

Error since RIFLEX 4.16.0.

Airfoil coefficients may now be given for a single Reynolds number. They will then be used for all Reynolds numbers.

Extrapolation outside the range of Reynolds values given will now be allowed. The value for the closest Reynolds number will be used; i.e. flat extrapolation.

The import of the airfoil library for foil load coefficients has been improved. This will reduce the analysis time for both static and dynamic analysis. The reduction will be most noticeable in static analysis.

The damping contribution to the radiation load was not included. This has been fixed. Both the added mass and damping terms now contribute to the radiation loads.

Error since RIFLEX 4.14.0.

Simo body system elements, i.e. Floater Forces Model Data, may now be used in linearized time domain simulations.

This is a new implementation and should be used with caution.

Corrected an error that caused restriction in the number of SIMO bodies that could interact hydrodynamically. The simulation failed if the number of bodies was larger than 3.

Corrected an error that caused simulations with dynamic current variation to fail unless the selected environment contained at least one irregular wave.

Long-standing error.

Currently, the FFT wave method must be selected in SIMO to allow wave kinematics at updated position to be used in RIFLEX for a coupled analysis.

The Main Riser Line (MRL) input allows the pressure to be specified at either end of the MRL. The specified inner pressure is, however, always applied at end 1 of the MRL. This will give errors in true wall tension and some stress components calculated in OUTMOD.

Long-standing error.

Specified nodal loads at nodes with user-defined skew boundary conditions are not handled correctly. The loads in the global directions are applied in the skew system. A warning is written to the .res file.

Nodes connected to a flex-joint will have degrees of freedom in a skew system. Specified loads at these nodes are not handled correctly. An error message will be written to the .res file and the analysis stopped.

Visualization is now enabled also for foil profiles not part of a wind turbine. The foil profiles have all to be described in a airfoil-library file.

A thickness correction based on a reference thickness and an exponent may now be included in the calculation of the fatigue capacity.

The calculated fatigue damage is now printed on the _outmod.mpf file.

The user may now specify an identifier for the fatigue calculations and may also specify all segments / elements / ends for fatigue calculations.

The criteria for skipping fatigue calculations and print at the first end of elements has been modified. Fatigue is now always calculated for all locations specified in OUTMOD. If fatigue is calculated for all elements with stored forces; i.e. NSECT = 0; fatigue calculation at the first end of an element is skipped if calculation was performed at the second end of the neighbouring element in the same segment.

Previously, fatigue was only calculated at the first end of the first element with fatigue calculation in each segment. As forces may not be stored for all elements within a segment, this may skip more nodes than intended.

The size of the integer and real input arrays in OUTMOD has previously been fixed to 1000 and 500 places. This is now increased to a minimum of 10000 and 6000 places and will be linearly increased if the OUTMOD work array size is larger than the default size.

This change decreases the available space for other arrays and may cause a previously successful OUTMOD run to fail. Increasing RIFLEX_OUTMOD_MEM by one (million Bytes); e.g. from the default 32 to 33; will solve this.

The coordinate system for the measurements is now based on the stress-free orientation of the shaft element where the electrical torque is applied. It was previously based on this element’s orientation at the end of static analysis. No change in results unless the shaft rotates around the global z-axis during the static analysis.

A minor change has been made to the additional nodal measurements for an external wind turbine controller that were added in RIFLEX 4.14.0. The nodal rotations are changed from the updated orientation of the neighbouring element to the rotations of the specified node from the stress-free to the current orientation.

The case that wave loads on elements are specified but no points for computation of wave kinematics are detected, will no longer lead to controlled program termination.

Transverse offsets may now also be specified for the second end of the last segment within a line type. See Transverse offset specification in INPMOD.

INPMOD.

Boundary change may be specified directly for a supernode. This is in addition to the existing boundary change for local segment nodes.

The Simo body system element is introduced to make it possible to acount for hydrodynamic interaction between bodies in a Riflex "coupled" analysis. The element nodes are of Simo body type, a new type of supernode. One Simo body is automatically attached to one simo body node having the same number of degrees of freedom as the Simo body. The Simo body system element may have one or more simo body nodes dependig how many bodies that interact.

The simo body supernode may be attaced to any already existing FE-node, or be used as master in a rigid super node connection.

The nodal forces can be given in a column format.

Rigid supernode connections may now be used in in linear time domain analysis

Visualization is enabled for linear time domain analysis.

A Bash shell script for running a complete coupled analysis has been added to the share/bin folder.

Error in system set up. Local line segment number used instead of global segment number for eccentric mass center; Cross sections of type CRS7.

This caused incorrect results if eccentric mass was specified for CRS7 cross-sections on lines other than the first line specified. The error has been corrected.

Error since 4.10.0

The combination of no wave forces acting on elements and export for visualization leads to uncontrolled program termination. The error has been corrected.

Include loads for nodal components at end 2 of a segment in linear time domain simulation. These were previously not included if the segment had more than one element.

Long-standing error

Neither Potential flow loads nor MacCamy-Fuchs loads are currently implemented for regular wave analysis. The is now checked and DYNMOD will give an error message and stop.

An alternative is to replace the regular wave with a tabulated ("numerical") wave spectrum with a single frequency component.

The default value for viscosity in air is set to 1.516E-5, air temperature 20 degree centigrades.

Error in static analysis if boundary change was specified during parameter variation. The error normally caused program termination, but could give error in results if the energy norm was used for solution accuracy.

Error since 4.12.0

Error if a node is specified to become a slave node during static parameter variation.

Long-standing error

Corrected an error in the initialization of the IEC 2005 wind events - extreme vertical wind shear, EWSV - extreme horizontal wind shear, EWSH - extreme operating gust, EOG - extreme direction change, EDC The wind events were incorrectly initialized when defined by a detailed specification; i.e. velocity or direction change and event duration. The error resulted in the event duration being zero and no change in wind.

Error since RIFLEX 4.12.0.

Corrected an error in the initialization of the IEC 2005 wind events for wind turbine class S. The expected turbulence intensity, Iref, was incorrectly set to the reference wind speed, Vref.

Error since RIFLEX 4.12.0.

Corrected an error in fatigue calculations for systems with only bar elements.

Long-standing error.

The echo on the _inpmod.res file of the Froude-Krylov scaling factor in the normal direction was incorrectly identified as in the tangential direction.

Note that the time domain VIV load model is currently restricted functionality and that a license is required.

Error since RIFLEX 4.14.0.

The time domain VIV loads may now be used with consistent mass / load formulation.

Note that the time domain VIV load model is currently restricted functionality and that a license is required.

Error since RIFLEX 4.14.0.

The option IUPPOS = 0, hydrodynamic loads calculated with lines kept fixed in static position and no update of surface penetration, is no longer available for analyses with SIMO bodies.

The default values for air and water density in the ENVIRONMENT CONSTANTS data group in INPMOD will be changed. The current default values of 1.3 and 1025 for air and water respectively are only applicable if the units `m' and `kg' are used.

In RIFLEX 4.18.0, the current default values will be used if `kg' is specified for the mass unit. Alternative default values will be used if `Mg' is specified. For other mass unit names, teh air and water density must be specified by thee user.

Two default values will be removed in the OUTMOD data group CODE CHECK CURVES in RIFLEX 4.18.0. - minimum yield stress, SMYS, and - modulus of elasticity, EMOD.

The current default values of 220.0E3 and 2.1E8 are only applicable if the units `m' and `kN' are used.

The CARISIMA riser-seafloor contact is deprecated and will be removed in the future.

The old, undocumented OUTMOD data group TIME FATIGUE LIFE will be removed following the 4.16 release.

The OUTMOD data group TIME FATIGUE DAMAGE should be used instead.

A scaling factor for the tangential Froude-Krylov load term has been added for Morison or potential flow loads. The scaling factor is available for the CRS0, CRS1, CRS2 and CRS7 cross-sections.

The option to specify no hydrodynamic loads has also been added for these cross-sections.

The input is extended in order to make the loads on bottom-fixed cylindrical monopiles to be applicable for floating single column systems by including a simple load model representing the radiation forces.

The (horizontal) radiation loads will be based on an added mass coefficient and a damping coefficient, AMY and DAMP.

In the vertical direction the wave excitation forces is calculated in a similar manner as the corresponding load term in the Morison equation. A user specified added mass coefficient will be input and used for the vertical direction, CAX.

Example of input (with default values):

Note that the extension of the MacCacmy & Fuchs input in Riflex 4.14.0 is not compatible with earlier versions.

Additional nodal or element responses may now be made available as additional measurements to an external wind turbine controller. This functionality is currently under implementation.

Irregular simulations with vessel transfer function NONE are now permitted. The vessel motions must be read from file or not be applied, i.e. the irregular motion indicator must be FILE or NONE.

The BEM code has been made less time-consuming. Automatic reshaping of matrices when calling subroutines have been avoided.

Improved memory handling to avoid program termination for analyses with large amount of input data.

Improved accuracy for 3D grids with very short distances between grid points. Improved data checking and warnings when reading the seafloor geometry file.

The water plane stiffness was zero if current was present and the element crossing the water line did not have attached nodal components.

Long-standing error.

Minor changes in convergence for cases with lumped formulation and current are expected.

Corrected a long-standing error for low frequency vessel motions from file. The specified low frequency motions were only applied if wave frequency motions were also specified.

Corrected the calculation of the "measurement" accumulated aerodynamic torsional load which is made available to external wind turbine controllers. Previously, the accumulated global x-moment was exported to the control system.

No error in applied aerodynamic loads.

Error since 4.12.0.

Currently, the FFT wave method must be selected in SIMO to allow wave kinematics at updated position to be used in RIFLEX for a coupled analysis.

The Main Riser Line (MRL) input allows the pressure to be specified at either end of the MRL. The specified inner pressure is, however, always applied at end 1 of the MRL. This will give errors in true wall tension and some stress components calculated in OUTMOD.

Long-standing error.

The MacCacmy & Fuchs input has been changed in RIFLEX 4.14.0. Old input must be updated.

Example of input (with default values):

Avoid overwriting the beginning of the array containing element forces and moments in nonlinear time domain simulation. The values for the first seven elements were temporarily zero during the time step. No consequence for normal analyses as the correct values were used in the calculation of geometric stiffness and in reporting.

The maximum number of DMS arrays is increased from 10503 to 20003 to allow larger models to be read by INPMOD. The INPMOD work array size is increased to 2 million integer words.

An error introduced in version 4.12.0 made it impossible to run generic external control systems in a RIFLEX Coupled model. This has now been resolved.

Restrict application of marine growth to CRS0, CRS1, CRS2. CRS3, CRS4 and CRS7 cross-sections. Volume loads and Morison hydrodynamic loads will be modified. The modifications are calculated with the assumption that the cross-section is circular.

The modification of Morison coefficients is now based on the increase in hydrodynamic diameter so that it is consistent with the initial non-dimensional coefficients.

Numerical errors and not-a-number ("NaN") results are avoided by skipping modification of the volume loads if the external area is zero.

Modification of the Morison coefficients is skipped if the initial hydrodynamic diameter is zero.

An error has been fixed where RIFLEX would crash when eigenvalues was calculated for a model with SIMO bodies (coupled model).

Error since 4.12.0.

The drag coefficients are corrected for marine growth.

The MacCamy-Fuchs loads are only calculated for the wind-sea part of a wave condition. An error exit is therefore added if a wave condition containing swell is specified for a case with MacCamy-Fuchs loads.

The application of marine growth in static analysis for systems with non-consistent units has been corrected. No error for analyses with consistent units; i.e. gcons = 1.0.

The error caused the volume loads of the whole system to be scaled by a factor of 1/gcons when growth was applied in static analysis. As the typical value of gcons for non-consistent units is 0.001, the error gave clearly non-physical results.

The error occurred even if a growth scaling factor of zero was specified.

Wave kinematics may now be calculated at updated dynamic positions in a coupled simulation. This resulted in an error exit from RIFLEX DYNMOD in 4.12.0.

Error corrections and enhanced funtionality to allow for swell and short-crested waves. The functionality is currently under testing.

Wind turbine response storage is available based on user input specification.

The drag force input specification has been expanded to allow for dimensional drag force coefficients.

A convergence criterion based on energy has been implemented in Stamod and Dynmod. The rotational degrees of freedom are included, as opposed to the existing displacement convergence criterion. The new convergence criterion is approximate because the unbalanced loads refer to the configuration at the previous equilibrium iteration. Hence, false convergence may occur for problems with oscillating convergence behavior. The energy convergence criterion can therefore not be applied without also using the displacement convergence criterion.

The blade root (at end 1 of the first element of the foil line) bending moments about the local x, y, z axes are now available to the external wind turbine controller.

The Morison drag term has been modified to be in accordance with the drag term used for axis symmetric cross sections. Please confer the user documentation.

The wave components of a realization of an irregular seastate may now have stochastic amplitudes. The new option CHAMP is used to specify deterministic (`DET') or stochastic (`STOCH') amplitudes.

Note that using stochastic wave amplitudes will cause the significant wave height to vary between realizations.

Wave kinematics may now be calculated at updated dynamic positions during a time domain simulation by specifying IUPPOS = 2.

Wave kinematics can be calculated during the simulation for ISURF ⇐4; i.e. potential fixed at mean water level (ISURF ⇐ 1 and ISURF = 4), potential stretched / compressed to instantaneous wave surface (ISURF = 2) and potential moved to instantaneous wave surface (ISURF = 3). Wave kinematics may not be calculated during the simulation for second order waves (ISURF = 5).

Except for very short simulations, this will be significantly slower than pre-generating the wave kinematics at the static position. This is because complete wave kinematics time series may be generated using FFT, while wave kinematics at individual time steps are calculated as a sum of all the non-zero wave components.

The option IUPPOS = -2; calculate wave kinematics at the static positions during the simulation; is added for verification and comparison.

Wave kinematics may now alternatively be stored on an additional file in ASCII format. Previously, only the binary format was available.

Wave kinematics calculated during a simulation; IUPPOS = 2 or IUPPOS = -2; are stored on files _updkin.asc / _updkin.bin.

When riflex.bat is called with CALL, variables such as PATH will also be changed in the calling environment. Multiple calls to riflex.bat will thus increase the length of the PATH variable, which may lead to the execution stopping if it becomes too long.

Calling riflex.bat with "CMD /C" will ensure that variables in the calling environment will not be changed. This is therefore recommended, e.g.:

Either backward slashes ( "\" ) or forward slashes ("/” ) may be used in the path to the riflex.bat file.

riflex.bat has been modified so that the PATH variable is now set back to its initial value after running RIFLEX.

Also included in RIFLEX 4.10.5.

The wind load was earlier activated together with specified forces. Now the wind load on turbine blades and and the rest of the structure may be activated as a separate load group in Stamod.

Example of input:

The numerical performance of the hysteresis bending model for CRS1 with IEJ=1 and IMF=1 has been improved. A new procedure for updating the hysteresis bending moments based on Backward-Euler integration was implemented, which allows for a fully consistent tangent stiffness matrix that is expected to improve the convergence properties of the solution procedure.

A duplicated node occurs when the last node of a line segment is the first node of the next segment. This shared node has one unique global node id but has two local node ids known to the user: lineid-segm1-lastnode and lineid-segm2-1.

Previously, only the data for the first of these were written to the ascii or binary storage file. The other was skipped. This occured even if both were requested part of the DISPLACEMENT RESPONSE STORAGE card. Now, both will be written - if requested.

Wind type 14, stationary uniform wind with shear, has been added. A power or logarithmic shear profile may be specified.

This wind type differs from wind type 10, stationary uniform wind with shear, values interpolated at grid points, in that the shear profile is used directly.

An extreme wind event from `IEC 61400-1 Wind turbines – Part 1: Design requirements – 2005'' may be applied in a time domain simualtion. The wind event can be applied to a stationary uniform wind with shear, IWITYP = 14, in both `RIFLEX analyses and in coupled RIFLEX - SIMO analyses. The following extreme wind events are available: - IEC 2005 extreme coherent gust with direction change, ECD - IEC 2005 extreme vertical wind shear, EWSV - IEC 2005 extreme horizontal wind shear, EWSH - IEC 2005 extreme operating gust, EOG - IEC 2005 extreme direction change, EDC

The EWSV and EWSH events can only be used for systems which include a RIFLEX wind turbine. The ECD, EOG and EDC events may be used for systems without a wind turbine, but some event parameters must then be specified; e.g. velocity change, direction change and duration of event for an ECD event.

Wind turbine shutdown with generator fault conditions may be applied in a time domain analysis. The shutdown event can be applied to a RIFLEX modelled wind turbine in both RIFLEX analyses and in coupled RIFLEX - SIMO analyses. The generator fault options include total loss of generator torque and bakup power formulated as following scaled torque control. In addition a mechanical brake in form of linear torque damping may be modelled.

Wind turbine blade pitch fault conditions may be applied in a time domain analysis. The fault conditions can be applied to a RIFLEX modelled wind turbine in both RIFLEX analyses and in coupled RIFLEX - SIMO analyses. The following blade pitch fault types may modelled: - Seized, i.e. fixed blade pitch from a specific time - Runaway, i.e. blade pitch change at a specific rate until reach of final pitch angle - Actuator bias, i.e. constant pitch deviation from required pitch

The blade pitch fault conditions are modelled individually for blades with pitch fault.

Pipe-in-pipe contact forces and moments in the global system are now stored on the file _cntfor.asc / _cntfor.bin along with the forces and moment in the local system. The contact forces are stored if element forces are stored on an additional ascii or binary file; i.e. IFORFM /= 0; and pipe-in-pipe elements are present in the system.

Wind speeds and aerodynamic forces acting on wind turbine blades are made available for visualization in SIMA and SimVis.

A test version of hydrodynamic loads based on potential theory (WAMIT) has been implemented. The functionality is currently under testing.

The marine growth is defined by specifying the thickness and density for a range of depths.

Example of input:

Marine growth is defined as a new static load group. During the load steps in this load group, the mass and diameter properties of the elements in this zone will be modified based on their current location. These properties will be fixed in the rest of the static analysis and in the dynamic analysis. The user may thus select the static configuration that is most representative for the conditions for which the marine growth is accumulated.

An overall scaling factor for the accumulated thickness will allow the user more easily to switch between no marine growth, partly accumulated marine growth and fully developed marine growth.

Example of input:

Riflex failed or gave incorrect results if the number of foil blades was different from 3.

Corrected in RIFLEX 4.10.4.

Corrected a long-standing error in hydrodynamic drag loads on CRS2 and CRS7 cross-sections when the lumped load formulation is chosen; i.e. LCONS = 0 in STAMOD.

Corrected in RIFLEX 4.10.5.

The input scaling factor for the Froude-Krilov term was not used for for CRS2 and CRS7 cross-sections if the consistent load formulation was chosen. Error in applied loads if - consistent load formulation was chosen; LCONS = 1 in STAMOD; and - CRS2 or CRS7 cross-sections with SCFK /= 1.0 were used.

The accumulated line length used in result presentation is now calculated from the actual stress-free segment lengths given in the line type definition in INPMOD; i.e. parameter SLGTH0 if it is given, otherwise SLGTH.

The line length is mainly used for results on the MatrixPlot files.

Previously, the line length was calculated from the updated stress-free element lengths in STAMOD and DYNMOD and from the final static nodal coordinates in VIVANA and OUTMOD.

The change in line length will be small for cases without - length changes due to temperature or pressure changes - element length modification for elements on / leaving a winch - large static elongation / compression

A minor error in the application of the torque from the wind turbine controller has been corrected. The torque was applied in the flex-joint local element system and not in the skew system at the flex-joint nodes. The difference between the systems is small and the correction gives insignificant changes to the results.

Corrected in RIFLEX 4.10.2.

A long-standing error (since version RIFLEX 4.4) has caused incorrect results or uncontrolled error termination when computing airfoil forces on cross-sections that were not part of a wind turbine. The error has been corrected.

Corrected in RIFLEX 4.10.2.

For coupled analyses with SIMO wind type IWITYP >= 10: - =10: Stationary uniform wind with shear - =11: Fluctuating uniform 2-component wind - =12: Fluctuating 3-component wind read from files (IECWind format) - =13: Fluctuating 3-component wind read from files (TurbSim format)

The wind at the updated coordinates of the RIFLEX hub supernode is now reported on the wind turbine result file. Previously, the wind at the SIMO body WIND_REF was reported. The reported wind is also sent into the external control system, but is otherwise not used in the analyses. The incoming wind on the blades is found using the updated nodal coordinates along the blades and has not been changed.

No changes for coupled analyses with SIMO wind type IWITYP < 10; the wind at the SIMO body WIND_REF is reported. Note that this wind is calculated at the wind force coefficient height ZCOEFF. This wind is also used as the incoming wind along the wind turbine blades .

Possible change in response for wind turbines with external control system and SIMO wind type IWITYP >= 10.

Corrected in RIFLEX 4.10.2.

Corrected an error that could occur when releasing a ball-joint in a nonlinear time domain simulation for a system that also contained one or more flex-joints. The error occurred if - a single, specified ball-joint was released and the system contained a flex-joint with the same reference number as the ball-joint - all ball-joints were released and the system contained any flex-joints

The error caused DYNMOD to terminate with an error message when the ball-joint connector was released. Error since flex-joints were introduced in RIFLEX 3.6.0.

Corrected in RIFLEX 4.10.2.

The export of hydrodynamic water depth to the .vtf file for animation in Xtract has been corrected.

Error in RIFLEX 4.10.0 and 4.10.1, corrected in RIFLEX 4.10.2.

The input of strain-dependent cross-sectional axial damping has been corrected. The table is given as IDMPAXI pairs of strain and damping coefficient values. All values are given on a single input line, as described in the User Manual. The table has previously been incorrectly read as two values on each input line.

This input may still be given as two values on each input line in RIFLEX 4.10.x. This may be chenged in later versions.

Corrected in RIFLEX 4.10.2.

The contact force file _cntfor.asc / _cntfor.bin was empty if seafloor contact elements were present. The contents were instead written to the Fortran file 80.

Error since RIFLEX 4.10.0. Also corrected in RIFLEX 4.10.4.

Several support vessels may now be specified without motion transfer function, i..e with IDWFTR = NONE. Previously, only one vessel in a system could be specified without motion transfer functions.

Corrected errors that could cause a dynamic simulation to fail during initialization of visualization. The error occurred if the simulation included several vessels and either - no vessel motions were included; CHMOT = NONE; or - vessel motions read from file; CHMOT = FILE; and several vessels did not have transfer functions.

Storage of wind turbine responses requires user input specification.

Since the input specification has been expanded to allow for dimensional drag force coefficients, an input code for hydrodynamic drag coefficients has to be specified. As a consequence the input is not back-compatible for MacCamy-Fuchs wave loads.

Static wind loads on RIFLEX elements will now only be applied in static analysis if WIND is specified as a static load. Previously, the static wind loads were activated together with specified forces. Add the load type WIND to the static load group with SFOR, specified forces, to get the same static loading as before.

See the section Wind force as separate load group in Stamod below for more deals.

The SIMO wind types that are not available in standalone RIFLEX; IWITYP < 10, are now only allowed in coupled analysis if wind loads are only applied to SIMO bodies and wind turbine blades. If the case includes non-blade elements with wind force coefficients IWITYP >= 10 is required.

Previously, the wind time series at the first SIMO body with wind coefficients was used for all non-blade elements with wind coefficients. The analysis stopped with an error if none of the SIMO bodies had wind coefficients.

Specified nodal loads at nodes with user-defined skew boundary conditions are not handled correctly. The loads in the global directions are applied in the skew system. A warning is written to the .res file.

Nodes connected to a flex-joint will have degrees of freedom in a skew system. Specified loads at these nodes are not handled correctly. An error message will be written to the .res file and the analysis stopped.

Generation of low frequency motions from spectra has been removed in RIFLEX 4.11.3. The data groups LFMOTION SPECTRUM SURGE, LFMOTION SPECTRUM SWAY and LFMOTION SPECTRUM YAW have been removed and the options CHLFM = GEN and CHLFM = NEW are no longer allowed in IRREGULAR RESPONSE ANALYSIS.

Time series of low frequency motions may be read from file.

The bottom tangent option for SB and AR systems, IBTANG, has been changed. IBTANG = 1 will now specify the 3D bottom formulation.

Note that the outer contact radius, R_EXTCNT, of the cross-section is used in the 3D bottom formulation while it was not used in the flat bottom formulation.

The 3D seafloor formulation gives contact on all nodes that are below Z < ZBOT + R_EXTCNT; i.e. contact at the outer contact radius. The original flat bottom formulation gives contact for nodes with Z < ZBOT; i.e. contact at the centreline. Insignificant changes are expected for cases in which R_EXTCNT = 0 for the segments in contact with the seafloor.

Results will change for cases where segments with R_EXTCNT > 0 had contact with the flat seafloor formulation. This may be handled in four alternative ways: - If the segments do not have other contact, set R_EXTCNT = 0 - If R_EXTCNT is the same for all segments with seafloor contact, lower the seafloor by R_EXTCNT. - Raise the final static coordinates of the nodes with specified position on the seafloor by R_EXTCNT. The total line length may have to be modified to obtain the same configuration and tension. - Continue using the original flat bottom formulation (see below), but note that this will not be available permanently.

The original flat bottom formulation may be chosen by specifying IBTANG = -9. This option has been made available to allow investigation of differences between the two formulations and will be removed in a later version.

Previously, the seafloor friction contribution to torsional load was activated in STAMOD by specifying the load type FRIT. This is now specified though the seafloor contact parameter ILTPR given in INPMOD. Specifying FRIT in STAMOD will now result in an error message.

The Carisima INPMOD data group NEW COMPonent SOIL has been renamed NEW COMPONENT SEAFloorcontact and will now be used for other seafloor contact types as well. The first line of input has therefore been split into two lines and a the parameter CHSFCT added to specify the seafloor contact component type.

Example of old Carisima input:

Example of new input:

An error has been corrected in the storage of element forces for cases with pipe-in-pipe contact. The error could lead to specified element curvatures not being stored as expected.

Storage of element forces is specified in the DYNMOD data group FORCE RESPONSE STORAGE. If IFORFM, Format code for storage and / or output of element forces, is not 0, the element forces for the specified elements are output to the additional file; _elmfor.asc or _elmfor.bin. If the model contains pipe-in-pipe contact elements, the contact forces for all pipe-in-pipe elements are written to the file _cntfor.asc or _cntfor.bin.

No error if forces were not stored on additional Ascii or binary files; i.e. if the data group FORCE RESPONSE STORAGE was not given or if IFILFM = 0. No error if the number of elements for which force storage was specified was at least 40% of the number of pipe-in-pipe elements.

Error since RIFLEX 3.6.

An error in the selection of airfoil characteristics for elements which are not part of a wind turbine blade has been corrected. In previous versions, the airfoil characteristics for complex lines could be assigned out of order in some cases.

The calculation of wind velocity could fail if all resulting wind components at a point and time were exactly zero. Scaling of the components resulted in a division by zero and an illegal number (NaN) was returned. This has been corrected.

Contact forces from all non-Carisima seafloor contact may be exported from STAMOD and DYNMOD for visualization in SimVis.

A new cross-section type CRS7 has been implemented which accounts for eccentric shear center, mass center and area center.

The CRS7 cross-section also employs a new element geometric stiffness matrix that accounts for the change of internal loads due to element rigid body rotation. This is expected to improve the Newton convergence properties and increase the maximum step size for simulations with large rigid body rotations and low tension.

A coupled bending and torsion model has been implemented for CRS0, CRS1, CRS2 and CRS7. The model employs a second order approximation of both the cross-section rotation and the longitudinal Green strain, and may therefore allow for increased element lengths in certain problems.

MacCamy-Fuchs loads (with additional quadratic drag) may be applied to vertical cross sections, which are assumed to be circular. The input format for CRS0, CRS1, and CRS2 cross sections is modified, existing files may be used without modification.

The 3D seafloor contact formulation has been improved in order to increase the numerical robustness. The modification will give minor changes to the results for some cases.

A new option has been added that allows the SIMO DP system to receive line tension measurements from RIFLEX lines in a coupled simulation. Previously this has only been possible for SIMO catenary lines. See the the SIMO userguide for more details.

An advanced aerodynamic option is included in order to remove the induction calculation on a wind turbine. This option is useful for a parked or idling wind turbine, where the aerodynamic loading is better described by quasi-static airfoil loads.

An aerodynamic option is included in order to user control of the Prandtl correction options. This is done by introducing on/off switches for correction at the blade tip, at the blade root, and a switch for how these correction factors are modified in yawed inflow.

The data group Seafloor contact specification may be given to specify which segments in an AR-system have contact with the seafloor. Different segments can have different spring-friction contact or have contact modelled using the new riser-soil contact formulation.

A single-node contact element is generated at each end of each beam or bar element in the specified segments.

The 2nd order wave calculations have been made more efficient. Slight, but not significant, changes in the results are expected.

The restriction limiting the number of short-crested directions to a maximum of 17 has been removed. This applies to both standalone RIFLEX analyses and to coupled RIFLEX - SIMO analyses.

The maximum number of line types is increased from 200 to 500 and the maximum number of components from 200 to 500.

The original flat bottom formulation will be removed in a later version. It may be specified with IBTANG = -9 in RIFLEX 4.10.x.

Generation of low frequency motions from spectra will be removed in a later version of RIFLEX. This functionality is currently available by setting CHLFM = GEN in IRREGULAR RESPONSE ANALYSIS and then giving the date groups LFMOTION SPECTRUM.

The simplified modelling of kill and choke lines attached to a tensioned riser by including them in the riser elements will be removed in a later version of RIFLEX. This functionality is currently available through the variable NAKC in ARBITRARY SYSTEM AR and then giving the input described in Description of kill and choke lines.

The sign of the applied pitching moment on airfoils which are not a part of a wind turbine was incorrect. This has been corrected.

Corrected in RIFLEX 4.8.4.

The record length of the binary files was set to four times the correct value. The .ffi, .sam, .raf, .bin and .ts files were therefore four time their necessary size. The .bin and .ts files were not comparable with their documentation and pre-existing tools for reading them.

Error since 4.8.0.

Corrected in RIFLEX 4.8.3.

An error in the dynamic implementation of flex-joint has been corrected. The error concerned flex-joints where the torsional rotation was free while the bending degrees of freedom were locked. This type of flex-joint is typically used in wind turbine modelling.

The error led to incorrect results if the local x-axis of the flex-joint did not coincide with the global x-axis. The error was negligible if the deviation from global x-axis was small. For larger deviations, the solution tended to diverge and the simulation stopped.

Consequences for wind turbine simulations where the flex-joint x-axis was close to aligned with the global x-axis: Except torsional moment, this error affects all reaction forces/moments in the first element in the shaft (between the hub and the flex-joint). Responses in the rest of the system are only slightly affected.

Corrected in RIFLEX 4.8.2.

The reading of the drag amplification factors from the specified MatrixPlot file has assumed that there were five lines between the line

and the first line with drag amplification values. This will not always be the case; the current version of VIVANA, for example, has four lines here.

The decoding now follows the description of the MatrixPlot file format in the MatrixPlot User Guide.

Corrected in RIFLEX 4.8.2.

An error in the 3D seafloor friction forces has been corrected. In the transformation of relative displacements and velocities from the global to local seafloor system the transformation matrix for the first node of an element was used for both nodes of the element. No error if the seafloor slope was the same at both ends of the element. The consequences of the error are expected to be insignificant for most cases.

Corrected in RIFLEX 4.8.1.

A possible error in selection of kinematics nodes for lines that cross the upper or lower limit for wave kinematics has been corrected. The error could cause incorrect kinematics at nodes with interpolated kinematics near the limit. Previously, the node with interpolated kinematics nearest a supernode without kinematics was selected as a kinematics node. This is now only done if there are no kinematics nodes between this node and the supernode. This correction may lead to less accurate kinematics. Please note that the selected kinematics nodes are printed on the *_dynmod.res file, so any change in behavior can be detected in this way.

Corrected in RIFLEX 4.8.0.

A long-standing error could give extrapolation in direction of the vessel motion transfer functions for short-crested wave components. The error occurred if the average propagation direction, WADR1 or WADR2, was less than

from the first or last direction that the transfer functions were given for. IWADR is the number of directions used in the spreading function, IWADR1 or IWADR2 in data group NEW IRREgular SEAState.

For IWADR = 5; the transfer function will be extrapolated if the average wave direction is within 60 degrees of the first or last direction the transfer function is given for.

For IWADR = 11; the transfer function will be extrapolated if the average wave direction is within 75 degrees of the first or last direction the transfer function is given for.

Corrected in RIFLEX 4.8.0.

Previously, the vessel motion transfer functions could be extrapolated in direction if the first and last transfer function directions did not cover a full circle; i.e. HEAD(NDHFTR) - HEAD(1) < 360 for ISYMHF = 0 or HEAD(1) not 0 for for ISYMHF > 0.

To avoid extrapolation, the first direction will now be repeated for the direction HEAD(1) + 360 if the specified directions do not cover a full circle.

Corrected in RIFLEX 4.8.0.

Irregular simulation without waves may now be run. This previously led to an error termination at the beginning of the analysis.

Corrected in RIFLEX 4.8.0.

Irregular simulation with low frequency motions and no wave frequency motions may now be run. This previously led to an error termination during generation of motion time series in DYNMOD.

Corrected in RIFLEX 4.8.0.

SIMA and SimVis have always visualized regular waves with zero degrees direction, regardless of the wave direction specified. This error has been corrected. The error has no consequences for other results.

Corrected in RIFLEX 4.8.0.

The functionality Nonlinear Buoyancy Correction (NBC) has been made available for coupled analysis. This was not the case in the RIFLEX version 4.6.

Corrected in RIFLEX 4.8.0.

The fish net load model requires that the net plane is defined. The plane is defined by the local element X-axis and a reference vector specified in the input group: LOCAL ELEMENT AXIS DEFINITION.

If the local reference vector was not given, an uncontrolled program termination occurred in STAMOD. The program will now check whether a reference vector is given for all fish net elements. In this case, the program terminates in a controlled way giving an appropriate error message.

Corrected in RIFLEX 4.8.0.

Several errors regarding the sheltered closed option for pipe-in-pipe contact have been corrected. This could cause errors in the calculation of the inner fluid load effects. Error in version 4.6 only.

Corrected in RIFLEX 4.8.0.

The input value of the seed for the phase angles of the different frequency components from VIVANA has previously not been used. This has been corrected in DYNMOD.

Corrected in RIFLEX 4.8.0.

Second order wave kinematics may now be used together with short-crested seas (irregular waves with directional spreading).

The tower shadow influence for upwind wind turbines is modified such that the drag coefficient and Bak correction factor may be taken into account. If these inputs are zero (default), the same tower shadow as before is used.

Downwind wind turbine modelling is enabled, including correct default blade orientation and control system actions. A cosine-squared type tower shadow for downwind wind turbines is implemented.

Morison-type aerodynamic drag forces may be applied to the dry part of CRS0, CRS1, CRS2, CRS3, and CRS4 elements.

Wind files generated by WASP Engineering’s IEC Turbulence Simulator can now also be read by RIFLEX built for Linux.

The new Linux release of SIMO, RIFLEX and VIVANA is 64-bit and solves several issues. Unfortunately, this means that 32-bit Linux operating systems are no longer supported.

The package has been tested on the following Linux distributions:

The friction stiffness along the pipe surface has been corrected. This is expected to improve convergence in dynamic analysis.

The unit of STIFFR, spring stiffness associated with the static friction coefficient, was incorrect in both the User Manual and the echo on the _inpmod.res file. The correct unit is F / L^2.

Corrected in RIFLEX 4.6.2.

An error that prevented the master pipe of a pipe-in-pipe pair to be a main riser line has been corrected. The error resulted in an controlled error termination before start of time integration when running dynamic analysis.

Corrected in RIFLEX 4.6.2.

Corrected an error that prevented more than one main riser line being defined. Error since 4.0.

Corrected an error in wave loading on RIFLEX elements from numerically defined (tabulated) spectra with multiple directions in coupled SIMO-RIFLEX simulations.

Corrected errors in calculation of tubular contact modelled using ELASTIC CONTACT SURFACE and NEW COMPONENT TUBULAR CONTACT. Changes in results are expected.

No chnage for pipe-in-pipe contact.

In both static and dynamic analysis, the stiffness terms along the pipe surface have been modified to improve convergence. The applied forces were not changed, so no significant change is expected in results with good convergence. Included in RIFLEX 4.6.2.

An option has been added to the pipe in pipe specification to allow the user to specify whether the inner pipe is exposed to the external environmental loads or is shielded by the outer pipe.

The default is that the inner pipe is exposed to the external environmental loads; i.e. buoyancy based on the water density given for the selected environment and wave loading based on the waves and current. Old input files will therefore give unchanged behaviour.

The pipes in a pipe-in-pipe pair may now be defined by main riser lines (MRL), lines or a combination of MRL and line.

Wind velocity at hub height in global coordinates is added to available measurements for the external wind turbine controller. Existing external wind turbine controllers can be used unchanged.

The internal initialization of dynamic stall parameters may fail for non-typical airfoils. The search methods have been made more robust, and error messages have been improved. The user is also given the option to specify the dynamic stall initialization parameters.

Alphanumeric identifiers may now be used for main riser lines (MRL). Old inputs with numbered main riser lines may still be used.

The simplified modelling of kill and choke lines attached to a tensioned riser by including them in a riser elements will be removed in the next version of RIFLEX. In RIFLEX 4.0 this functionality is available through the variable NAKC in ARBITRARY SYSTEM AR and then giving the data in B6.9.

Generation of low frequency motions from spectra will be removed in the next version of RIFLEX. In RIFLEX 4.0 this functionality is available by setting CHLFM = GEN in IRREGULAR RESPONSE ANALYSIS and then giving the date groups LFMOTION SPECTRUM.

The restricted functionality to import internal flow data from file will be removed in the next version of RIFLEX. In RIFLEX 4.4 this functionality is available for some users by setting INDINT=2 in data group E1.4 and then giving the data group IMPORT FLOW DATA.

The motion period must now be the same for all vessels.

If no regular wave is specified, regular motions must be given in a regular analysis. If the system contained multiple vessels, the motions were incorrect for vessels 2 - NVES. This is now corrected. Note that the motion period must be the same for all vessels.

Corrected in RIFLEX 4.4.0.

The OUTMOD data group TIMEDOMAIN FATIGUE DAMAGE has previously reported zero fatigue damage if the stresses exceeded the stress level corresponding to failure after a single cycle. This is now handled as an error. The most probable cause of this occurring is that the units of the S-N curve and the calculated stresses are not compatible.

Also corrected in RIFLEX 4.4.2.

The expression for radius of gyration that is used for the moment of inertia around the local X-axis has been corrected. This concerns cross section type CRS0 and also cross sections created based on a stress joint specification. Further, the coating contribution to moment of inertia was not previously included. The radius of gyration has been calculated to be 71% of the correct value without coating. For cross sections with coating, the difference is larger. For normal riser analysis, the error had minor or insignificant influence on the results. However, this is system-dependent. The error has been corrected and the updated expression also accounts for the coating contribution to the moment of inertia.

Corrected in RIFLEX 4.4.1.

Corrected an error in the storage of dynamic analysis animation for DeepC. Error in 4.4.0 only.

Corrected in RIFLEX 4.4.1.

Corrected an error in the specification of explicitly selected nodes for undisturbed wave kinematics. This may be done by specifying DIFF combined with IVES = 0 in the DYNMOD data group IRREgular WAVE PROCedure. Error in 4.4.0 only, where an error exit was caused.

Corrected in RIFLEX 4.4.1.

An error caused OUTMOD to fail if the OUTMOD data group SUPPf TIME SERIes was given more than once. The error has been corrected.

Corrected in RIFLEX 4.4.2.

Blade pitch results are now presented for simulations which include an external control system.

Improved in RIFLEX 4.4.1.

The factor adjusting initial stiffness for the friction moment functionality has been added as user input. See Stiffness properties classification in INPMOD.

The line type definition may now include specification of a curved stress-free configuration. See Transverse offset specification in INPMOD.

TurbSim 3D wind files may now be read by RIFLEX. See Fluctuating 3-component wind field read from TurbSim file.

Static airfoil forces may now be applied to CRS2 cross sections that are not blades in a wind turbine.

A new option to read in and use wave kinematics from file has been added, see Additional detailed specification of wave kinematics points (optional).

The SIMO DP system is now called only at every SIMO time increment (Sampling time interval). Previously, the old SIMO DP system was called at each time step (Time integration step) while the new DP system was correctly called at the sampling time interval.

The correction will possibly increase the computational robustness when the old DP-system (marked deprecated) is applied.

Improved selection and presentation of results for wind turbines. The results are now presented for all blades in the non-rotating shaft system.

Blank space is added around some numbers printed on the .res files. This has been done to increase readability and to facilitate comparison of numerical results.

Minor improvements in error handling and layout.

Windows versions 4.4 and higher are 64 bit executables and therefore require different Fortran and Java DLLs than earlier version. The necessary DLLs are included in the download package. The preformance is improved.

The default sampling step of the pre-generated time series, DTGEN, is decreased to 0.5 s. The default length of the pre-generated time series, TIMGEN, is increased to 4.6 hours. The default simulation length for irregular analysis, TIME, is increased to 3.06 hours.

The default value of the DYNMOD parameter IVARST, Code for automatic subdivision of time step, is changed from 2 to 0. If a simulation fails to run, check if it previously used subdivision and if the default value of IVARST was used.

Fixed a long-standing bug in pipe-in-pipe contact search that caused the search to fail in the vicinity of flex joints and ball joints. Users with pipe in pipe models that contains flex joints and/or ball joints should consider rerunning the simulations.

Corrected in RIFLEX 4.2.1.

Error in hydrodynamic drag forces on slender elements

If depth dependent current was applied and wind also was specified in the environmental condition, the computed drag forces on slender elements were in error. This was due to an incorrect z-position being used to interpolate the shear profile. The error has been present since version 4.2.0. The consequence is incorrect drag forces on the slender elements. Analysis done with this combination by version 4.2.0 should be re-run! The error has been corrected.

For other SIMO bugfixes, please refer to the SIMO Release Notes.

Corrected in RIFLEX 4.2.1.

Simulations occasionally failed when vessel motions were imported from external file. The algorithm for prescribed (vessel) rotation induced translation for nodes attached to the vessel has been modified to have improved numerical robustness. No or insignificant changes to results.

Corrected in RIFLEX 4.2.0.

An error in the exported vessel velocity and accelerations for coupled analysis with both SIMO bodies and RIFLEX support vessel was corrected. No error found in analysis results. No error for standalone RIFLEX or for coupled analyses with only SIMO bodies,

Corrected in RIFLEX 4.2.0.

A long-standing error in the winch formulation has been detected and corrected. Analysis based on previous versions of RIFLEX should be re-run.

Corrected in RIFLEX 4.2.0.

Linear drag coefficients have not been applied in linearized time domain analysis. Linear drag coefficients are seldom used, so the consequences are assumed to be minimal. The error has been corrected.

Corrected in RIFLEX 4.2.0.

Error for combined regular wave loading with specified harmonic displacements (IMOTD = 2). The motions after the first wave period were incorrect if a different period was specified for the motions than the wave period (and the wave period was mot a multiple of the motion period). . Error since RIFLEX 3.6.7.

No error if motions are calculated from vessel transfer functions (IMOTD = 1).

Corrected in RIFLEX 4.2.0.

New modelling features have been added to model transverse contact between a vertical pipe and the soil. The new INPMOD data group GEO SPRING SPECIFICATIONS is used to give the input to the model.

Extend the generalized Morison’s equation by optionally scaling the term associated with the Froude-Kriloff load. This to enhance the area of application of Morison’s equation, i.e. allow for larger structural diameter versus wave length. Note that only loads normal to the principal axis of an element are scaled. Additional, optional input is added for cross section typed CRS0, CRS1, CRS2, CRS3 and CRS4.

2nd order wave kinematics for long-crested waves have been added to RIFLEX. 2nd order wave kinematics are obtained by setting the DYNMOD input parameter ICOSIM to 5.

Generated wave kinematics may now be stored on additional binary file from DYNMOD. Data group IRREGULAR KINEMATICS STORAGE to store wave kinematics on an additional file <prefix>_wavkin.bin. The contents are described in the text file key_<prefix>_wavkin.txt.

The specified control interval for internal wind turbine control system has been activated. In addition the value of the integrator gain are kept within saturation limit. The reason for this implementation is to improve the behaviour of the internal control system.

Wind turbine results may now be stored on additional text file or binary file <prefix>_witurb.asc or <prefix>_witurb.bin. The contents are described in the text file key_<prefix>_witurb.txt. There is no special input for storing the wind turbine results. The results storage is based on storage specification for element forces, i.e. time interval for sampling and file format.

Allow NODSTP = 0 i.e. no kinematics for this line in the detailed kinematic specification.

Available from RIFLEX 4.2.1.

The included Java folder and the HLALIB.jar file have been updated in the Windows installation .zip file.