Slamming of large high-speed catamarans in irregular seas

Current design methods are limited in their ability to predict long term loading statistics relating to wave loads and fatigue from prolonged cyclic loading. These methods either neglect slam loads entirely or they are included as post-processed or simplified two-dimensional methods. The work presented in this thesis introduces a combined theoretical-empirical approach to determining long term load trends in realistic sea conditions during the initial stages of the design spiral. This method builds on a previously developed non-linear time-domain seakeeping strip theory method, for high-speed multi-hull vessels, using scale model testing in irregular seas as a basis for an empirical slam module. Towing tank experiments, using an extensively instrumented 2.5m hydroelastic scale model wave-piercing catamaran representative of the 112m class Incat design, were used to develop a database of slam events in a range of realistic (but idealised) irregular sea conditions. A total of 2,103 slam events were identified over 22 test conditions during the scale model experiments. Large slams generally occurred in the conditions where motions were largest; however significant scatter was present with extreme events observed to be up to four times the median for most conditions. Occurrence rates were found to be a function of encountered wave frequency and significant wave height. If the encountered wave frequency coincides with the motion resonance, slam rates increased. Increasing the significant wave height also increased slam occurrences. A wave height dependent slam threshold was identified by extrapolating occurrence rate trends with decreasing significant wave heights. Pressure measurements also revealed that the cross-deck structure was exposed to large local pressures at each measurement station, suggesting that ship designers should ensure the structure can withstand large local loads along the entire length of the bow. The non-linear time-domain seakeeping program was extended to simulate motions and loads in irregular seas and a method for constructing idealised wave spectra was developed as an input to the seakeeping code. The extended code was verified by conducting a series of program tests and then validated by comparing computational ship motions with results from the scale model experiments in the absence of slamming. A new module for predicting slam loads, based on a statistical analysis of scale model tests was then developed and integrated into the extended time-domain seakeeping method, allowing for the slam events to be determined on-line in the time-domain. Slams are identified by defining a location dependant immersion threshold criterion based on the geometry of the hull form combined with a stochastically determined variation originating from experimental observations. In the event of a slam, the maximum load and slam duration are determined by empirical methods stemming from regression analyses on experimental data. Vessel forward speed and relative vertical velocity at the centre bow truncation are used to predict the maximum slam load. Slams loads are ramped up and down over a number of time steps. Forward speed and immersion at the centre bow truncation are key explanatory variables in the duration calculation. The slam load is then applied over a number of time steps according to its duration at the location where it was first triggered. A case study is finally undertaken demonstrating the application of this method. A highspeed ferry service route was selected and real wave data used to determine expected wave environments. The computer simulation was run for a range of sea conditions and slam events are identified. Slam event statistics are then extrapolated to produce expectations for long term (20+ years) loading expectations showing how this method could be a valuable tool when considering the long term loading implications of a vessel for a particular route.