A predictor-corrector integration scheme for efficient time-domain simulation of floating wind turbines

Time-domain fully coupled simulation is essential for analysis and design of floating wind turbines (FWTs), yet its high computational cost remains a bottleneck for large-scale parametric and optimization studies. Conventional integrators such as the fourth-order Runge-Kutta (RK4) method require multiple force evaluations per step, which dominate runtime in aero-hydro-servo-elastic models. To overcome this challenge, this study develops a predictor-corrector one-force (PC-1F) scheme within the in-house simulation tool Pywind. The method enforces a single synchronized force evaluation per step while preserving second-order accuracy and eliminating empirical damping. Verification is conducted on the IEA 15-MW semi-submersible FWT under both the wind only, wave only and combined load cases. Results show strong agreement with RK4, with correlation coefficients above 0.95, mean discrepancy within 0.5% for key responses, and consistency of power spectral density in the wave-frequency band (0.05-0.3 Hz). Efficiency tests demonstrate iteration rates 70-90% higher than RK4, confirming that PC-1F achieves comparable fidelity at substantially reduced cost. These findings highlight PC-1F as a practical, engineering-ready alternative to multi-stage schemes. Moreover, the method offers promising applications in accelerated parametric sweeps, uncertainty quantification, and integration into multi-objective optimization frameworks for next-generation FWT design.