The Runge-Kutta method was employed to calculate the equations of motion. Input data for the simulation are the measured time histories of both rudder angle and crew incline angle at every 0.1 seconds. At the sea test, the boat ran straight course and circular maneuvering course in the calm water.

　

 Figure 8 shows the comparison between measured and simulated results during 15 seconds at the straight course running. In the figures, (a) shows time histories of rudder angle and crew incline angle, which are the input data for the simulation. In this case, although the boat runs in the straight course, it can be seen that the crew incline angle moves considerably. In order to compensate the slight positive roll angle of the boat in Figure (b), the crew is inclining from 0 to -10 degrees. Then (b) shows that the boat is stable in roll and able to keep upright state. Figure (c) shows the trajectory of the boat at every 1 second. The grid spacing is 30 meters. In this case, although the measured trajectory shows almost straight course, the simulated one shows slight drift to the starboard side. This drift in the simulation was caused by the shift of -1 to -2 degrees in rudder angle, which might be resulted by the slight misalignment between the rudder shaft and the potentiometer.

　

　

 Figure 9 shows the result during 40 seconds at the circular course running in left turn. In Figure (a), the rudder is steered around 5 degrees and the crew inclines from 0 to -10 degrees to the turning side. Figure (b) indicates the boat inclines around -7 degrees to the turning side too. The simulated roll angle coincides with the measured one until the boat starts to turn at 5 seconds. However, after the boat reaches the steady turning condition, the simulated roll angle becomes smaller than the measured one. This might be caused by the overestimation of stability in the turning motion, which was calculated on the assumption that the acting point of main hydrofoil was on the 20% point from each tip. However in Figure (c), the simulated trajectory coincides well with the measured one.