Abstract
This report presents a recommended framework for establishing operational criteria for dynamically positioned (DP) vessels, transitioning from single-statistic acceptance thresholds toward risk-adjusted probability-of-non-exceedance criteria. The methodology is demonstrated through two representative marine operations: (i) station-keeping for a semi-submersible drilling unit, and (ii) relative-motion response during gangway operations between a floatel and a turret-moored FPSO.
The proposed approach uses time-domain simulation models that capture first-order wave excitation, second-order (slow-drift) wave loads, turbulent wind, and steady current, along with realistic DP control logic that includes thruster dynamics. We recommend tuning these models against model-test data to capture the damping and stiffness characteristics.
Central to the approach is the Probability of Success (PoS) metric, with the target probability set to reflect the consequences of exceedance. Ensembles of simulations across environmental conditions yield extreme-value distributions of the maximum response, from which the probability of non-exceedance for critical limit states is quantified. Common descriptors of the maximum response — the Most Probable Maximum (MPM, the mode of the extreme-value distribution) and the Expected Maximum (𝐸max, its mean) — are not, in themselves, conservative acceptance criteria. For a Gumbel-distributed maximum response, the MPM is exceeded in approximately 63% of realisations and 𝐸max in approximately 43%; therefore, a high quantile such as 𝑃90 or 𝑃99, or an explicit PoS target, is preferred for risk-critical decisions.
The two case studies illustrate fundamentally different exceedance mechanisms: the drilling rig is capacity-limited, governed by thrust margin against environmental forces, whereas the gangway operation is precision-limited, sensitive to lateral and vertical relative motions that first-order waves readily excite. The report provides procedural guidance for operators, consulting engineers, and research bodies, including recommendations for selecting environmental test cases from hindcast data, defining simulation durations to ensure statistical reliability, and managing weather-forecast uncertainty through parametric, spectral, and ensemble-spectral approaches that resolve heading- and composition-dependent effects beyond a scalar 𝛼-factor.