Quantifying extreme ship loads and motions: statistical challenges and approaches
Abstract
Quantifying extreme (primary and secondary) loads and motions of a ship is of paramountimportance to the Navy as these often carry tremendous costs, both human and material.Indeed, extreme loads need to be considered at the ship design and building stages, andif not properly accounted for, can lead to structural damage. Likewise, extreme motionscan lead to a cargo or personnel loss, or even capsizing. The ever-changing nature of theNavy ship designs, in particular, calls for special cautions, as these designs can be prone tounexpected manifestations of extreme loads and motions. Understanding these extremes willthus increase the Navy~s readiness to counter weather-related and other risks, as it continuesupgrading its fleet, to meet the needs and challenges of the 21st century.Despite the critical roles played by extreme ship loads and motions though, their quantificationand understanding has lacked in fundamental ways. For example, it has becomepart of the folklore to fit the Weibull distribution to the observed loads or motions (or theirpeaks) and to use the fitted distribution to extrapolate into the tail regions of interest. But thispractice remains somewhat unjustified. The study of extremes is further complicated by thefacts that they are rare by their nature, and that they are often driven by non-linear aspectsof the studied phenomenon. For example, extreme loads result from the combination of anon-linearly changing stiffness, determined by the shape of the ship~s hull, and the form ofthe random incident wave.The main goal of the project is to provide a comprehensive understanding of the occurrencesof extreme ship loads and motions, together with justifiable, tested and validatedtechnical approaches to quantify them. At the center of this effort is the development ofa statistical model for the distribution tail of a quantity of interest (vertical bending moment,a particular motion, etc.), that captures the physics of the studied phenomenon, that is generalenough to encompass the various ship hull forms and sea conditions, and that is amenable todata fitting with practically acceptable statistical uncertainty (due to limited data). The fitteddistribution tail would then be used in quantifying the occurrences of extreme ships loads andmotions, from a limited data set of these in suitable numerical or basin simulations.Reaching the main goal is envisioned along the four fundamental and interconnected directions:A. Probabilistic and analytic methods for loads and motions and, in particular, ofthe resulting loads and motions distribution tails. B. Statistical analysis of loads and motionsdata from numerical simulations of varying degree of sophistication (SimpleCode, LAMP,etc.) and, in particular, their distribution tails. C. Development of a statistical model for distributiontails, accompanied by statistical fitting and inference procedures for data of limitedlength. D. Validation of the suggested quantification schemes for extreme loads and motions,across different sea conditions, ship geometries, operational parameters and simulationapproaches; Understanding and quantifying the model uncertainty related to the simulationschemes, in particular.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 24, 2019
- Source ID
- N000141912092
Entities
People
- Vladas Pipiras
Organizations
- Office of Naval Research
- United States Navy
- University of North Carolina at Chapel Hill