A new fatigue prediction model for marine structures subject to wide band stress process

• We propose a new probability density function which is composed of one half Gaussian and two different Rayleigh distribution function.
• Procedure of numerical formulation is introduced in detail.
• We assume various artificial wide band spectra from which stress process is deployed in time domain.
• Accuracy and applicability of the proposed probability density function is verified by comparing fatigue damage for various wide band spectra.

Marine structures on offshore fields are continuously subjected to wide band random loadings due to the combinations of wind-sea and swell, wind-sea and hull girder vibration, and wind-sea and drift/restoring. Those kinds of loadings are translated into wide band spectra with multiple peaks in frequency domain which violate a primary assumption of narrow band Gaussian process. It is known that it follows no more Rayleigh distribution and spectral fatigue procedure based on Rayleigh distribution therefore becomes inaccurate and conservative. This paper proposes a new probability density function (PDF) instead of Rayleigh PDF in order to improve fatigue prediction under wide band Gaussian process and explain a detailed development procedure. The validity of the new fatigue prediction model is first verified by comparing rainflow stress range distribution with two different PDFs: Dirlik and new ones. Then accuracy of the new model is second verified by comparing fatigue damages from various fatigue models with rainflow fatigue damages which is obtained from bimodal stress spectra.