Optimal maintenance support system for offshore structures based on wave prediction and motion analysis with NOWPHAS

The framework of a “wind farm” as an offshore structure with multiple wind power generation units is introduced. In Japan, these offshore structures are designed as floating structures owing to the bathymetry of the deep seafloor. They are situated in open seas, such as the Pacific Ocean or the Sea of Japan, where strong winds are frequent. However, these floating structures are occasionally damaged in rough seas. This can lead to the wind power generation being interrupted for a longer time period. This implies that maintenance and repair work needs to be performed regularly and quickly by specialist engineers. This study aimed to minimize the duration of wind power interruptions and examine various aspects. First, the wave height, wave period, and wave direction were predicted using a machine learning method with the coastal wave database in Japan as training data. Second, the movements of the working ship were reproduced based on the predicted results of the wave parameters to evaluate the safety and practicability of the engineers’ transfer from the working ship to the offshore structure for maintenance operations. Finally, an optimal maintenance support system for Japanese offshore structures was devised based on the study results. •Constructing machine learning to predict wave information one week in advance.•Estimating the ratio of significant wave heights below 1.5 m as an indication of calm seas.•The motion response of the working ship is calculated based on predicted wave information.•Safety criteria is different if the feasibility of transfer is evaluated by surge and sway motions.•Number of consecutive calm days is estimated and compared to evaluate maintenance scheduling.