Fig。 11 (top) presents the wave train of the so called New Year Wave  recorded in the North Sea and  simulated in the wave tank using the above introduced modified approach。 This rogue wave with an unusual Hmax/Hs ratio of 2。15 is applied to both the experimental and nu- merical investigation of rogue wave impact on the struc- tural (splitting) forces of a semisubmersible (Fig。 11, Clauss et al。 (2003b))。 Wave sequences calculated by the modified non-linear approach have also been  suc-

New Year Wave

200 250 300 350 400 450 500

corresponding splitting force

Fig。 11: Top: comparison of recorded New Year Wave and wave tank simulation (scale 1:81)。 Bottom: mea- sured and calculated splitting forces of the semisub- mersible due to the rogue wave impact (all data pre- sented as full scale data)。

cessfully applied to the investigation of rogue wave im- pacts on the vertical bending moments of a stationary crane vessel (Clauss et al。 (2003a)) and an FPSO ship (Clauss et al。 (2004b))。

Experimental investigation of intact sta- bility

For the experimental investigation of intact stability with regard to both extreme and resonance phenom- ena the wave train as the beginning of the cause re- action chain can be directly compared to the reaction of the cruising ship since all time series are calculated resp。 measured in the moving reference frame of the ship model。

For controlled capsizing tests (Clauss and Hennig (2003)) we generate a regular wave with an embedded ”Three Sisters wave” sequence at a moving reference frame。 Fig。 12 shows a wave packet within a regular wave measured at a stationary wave probe close to the wave board (x = 297。8 m, model scale 1:34)。 It is trans- formed to the position of the cruising ship。 As shown in Fig。 12 this resulting wave sequence is quite regular and contains the target ”Three Sisters wave” at the lo- cation of interaction with the cruising ship。 Thus, high roll angles (lower diagram) can be induced by generat- ing tailored moving reference frame wave trains deter- ministically。

Numerical predictions can also be directly compared to model tests applying the following scheme: The wave train used in the numerical simulation for assessing ship safety is given as full scale target wave train (Fig。 13 top) and transformed to model scale (1:34)。 Now the modified non-linear approach is applied to obtain the wave train at the position of  the  wave  maker。 Thus the corresponding control signal for driving the wave maker (signals for upper and main flap of double flap wave maker at Hamburg Ship Model Basin)。 The gener- ated wave train is registered at a stationary wave probe close to the wave maker and transformed to x = 125 m (compare target wave train)。 The ship position is mea- sured during the test。 Thus, the stationary wave  train is transformed to the moving reference frame of the ship

Fig。  12:  Roll motion of a multipurpose vessel (GM   =

1。44m, v = 14。8 kn und µ = ±20◦) in a regular wave from astern (λ = 159。5 m, ζcrest = 5。8 m) with proceed- ing high transient wave packet (compare Fig。 7)。

model to obtain the wave as experienced by the ship。 The resulting (measured) roll motion can be directly compared to this wave train (Fig。 13 bottom)。  The same test data is used for a visual comparison of numer- ical simulation and model test results in Cramer et al。 (2004)。

Discussion, conclusions and perspective

In this paper a modified non-linear approach for mod- elling wave propagation is presented which provides