Session: 02-07-01 Reliability of Marine Structures and Components

Paper Number: 130321

130321 - A Monte Carlo Based Model for Estimating the Reliability of Offshore Structures Exposed to Environmental Loading 

The load environment around offshore structures is complex. Wind, waves and ocean currents delivers complex time-varying load scenarios. With extreme waves ultimately breaking, transient load effects can be large, resulting in significant multi-DOF dynamic response. The resulting load effects may depend upon the structural system and the geometric location of a critical failure mechanism. Ultimately, in an optimized offshore structure, many failure mechanisms will contribute to the overall reliability of the structure. This interaction may be complex as different load-events may drive design loads across different failure mechanisms. The overall structural reliability will depend upon this complex correlation.

Historically, reliability models for offshore structures are often notional, as a range of often crude assumptions are needed to simply to allow use of available statistical methods, as e.g. FORM. It has e.g. been common practice to assume wave loading is well represented by (a deterministic) load calculation using regular wave theory. With latest knowledge proving the largest (and governing) ocean waves can be breaking wave events, regular (non-breaking) load effects may largely underestimate the true load effects across a range of failure mechanisms relevant for offshore structures.

In the last decade, work has been undertaken to develop a calibrated and as close to actuarial risk as possible SRA model with the aim to estimate structural failure probability for structures operated by TotalEnergies in the Danish North Sea sector. This complex calibrated model will be described.

The key engine in the SRA model is based on a crude Monte Carlo scheme. A Monte Carlo set-up was adopted to allow a full representation of the complex interaction between environmental loading from wind, waves and current and the stochastic multiple failure mechanism resistance of the structures.

Work includes development of a long-term statistical model of wind, waves, tide, surge and current. Hindcast models are calibrated/validated against the more than 30 year long record of environmental measurements in the Danish sector. Based on the calibrated/validated hindcasts the statistical model is developed and used for providing a record of 1h storm parameters in a period of 30 mill. years. The long-term record is based on 1h average of significant wave height, mean wave direction, wind speed, wind direction, current speed, current direction, surge, tide, etc. This approach allows correlation between individual parameters, e.g. the wind-direction will not be fully correlated with the mean wave direction, etc.

Short-term models are developed to expand the long-term 1h data into high resolution timeseries of loading. A 2-year wave-basin test program allowed calibration of a semi-empirical random phase sea state model (incl. breaking). A fully transient wind model, with correlation to occurrence of large wave events, was developed to allow realization of the fully transient wind load environment. The short-term models delivered stochastic timeseries of loads in each of the (millions) of 1h intervals in the long-term file.

A multi failure mechanism dynamic response model was developed to assess the load effects in each failure mechanism in the structure. The model was calibrated against a series of several hundreds of non-linear pushover analyses of the structure being assessed. The response model included stochastic variability of all resistance parameters like strength equations, yield stress, soil-stiffness, soil-strength, etc.

Ultimately, using a cloud-based set-up, the load effect in each failure mechanism was calculated through time domain analysis of each 1h sea states in the long-term file. Maximum load effects were compared to realizations of the failure mechanism capacity. The overall result being failure probability of individual failure mechanisms, and properly accounting for correlations, ultimately the structures (system) failure probability

Presenting Author: Allan Zeeberg TotalEnergies, TEPDK

Presenting Author Biography: Allan Rod Zeeberg is a Senior Structural Engineer with a career spanning since 2002, primarily in the domain of structural engineering for leading players in the oil and gas industry, specifically Maersk Oil & Gas A/S and TotalEnergies Denmark. The work has primarily revolved around the critical disciplines of structural load and resistance.

In the early phase of the career, encompassing the first decade, the focus was on design and re-assessment of structures, ensuring the integrity and safety of offshore installations. Responsibilities entailed overseeing the entire life cycle of structures, from the initial design phase through fabrication, transportation, and installation.

For the next decade, the career trajectory changed more into the territory of extreme loads and resistance, particularly driven by the extreme forces of breaking waves. This involved integration of new knowledge and innovative techniques to enhance structural resilience against the power of the sea. This involved not only the structural discipline but also the MetOcean discipline.

From the foundation of extensive wave basin model testing, cutting-edge tools and techniques for assessing the probability of structural collapse/failure including breaking waves was developed. This involved development of tools for full Monte Carlo simulation of full transient dynamic wave load calculations and structural response as well as a probabilistic approach for assessing the structural resistance. In addition, a deterministic approach was developed. Part of the work has been published (mainly within the MetOcean discipline) and further publication of the work is ongoing.

Authors:

Allan Rod Zeeberg TotalEnergies, TEPDK
Jesper Tychsen TotalEnergies, TEPDK
Jørgen S. Nielsen TotalEnergies. TEPDK
Jean-Marc Cholley TotalEnergies, HQ