Auto-control model building using machine learning regression for extreme response prediction

2020 ◽  
Vol 60 (1) ◽  
pp. 155
Author(s):  
Darrell Leong ◽  
Anand Bahuguni
Keyword(s):  
Machine Learning ◽  
Model Building ◽  
Response Prediction ◽  
Offshore Structures ◽  
Practical Implementation ◽  
Monte Carlo Data ◽  
Sea State ◽  
Extreme Load ◽  
Extreme Response

The long-term forecast of extreme response presents a daunting practical problem for offshore structures. These installations are subject to varying sea conditions, which amplify the need to account for the uncertainties of wave heights and periods across a given sea state. Analysis of each sea state involves numerically intensive non-linear dynamic analysis, leading to massive computational expense across the environmental scatter diagram. Recent research has proposed several effective solutions to predict long-term extreme responses, but not without drawbacks, such as the limitation to specific failure locations and the absence of error estimates. This paper explores the practical implementation of control variates as an efficiency enhancing post-processing technique. The model building framework exhibits the advantage of being fully defined from existing simulation results, without the need for external inputs to set up a control experiment. A composite machine learning regression model is developed and investigated for performance in correlating against Monte Carlo data. The sampling methodology presented possesses a crucial advantage of being independent of failure characteristics, allowing for the concurrent extreme response analyses of multiple components across the global structure without the need for re-analysis. The approach is applied on a simulated floating production storage and offloading unit in a site located in the hurricane-prone Gulf of Mexico, vulnerable to heavy-tailed extreme load events.

Estimating Extreme Response of Drag Dominated Offshore Structures From Simulated Time Series of Structural Response

2007 ◽  
Author(s):  
A. Naess ◽  
O. Gaidai ◽  
S. Haver
Keyword(s):  
Response Time ◽  
Structural Response ◽  
Weather Conditions ◽  
Response Prediction ◽  
Offshore Structures ◽  
Rate Function ◽  
Extreme Response ◽  
Time Histories ◽  
Short Time

The paper presents a study of extreme response statistics of drag dominated offshore structures, showing a pronounced dynamic behaviour when subjected to harsh weather conditions. The key quantity for extreme response prediction is the mean up-crossing rate function, which can be simply extracted from simulated stationary response time histories. Present practise for obtaining adequate extremes for design purposes requires a number — say 20 or more — of 3-hour time domain analyses for several extreme sea states. For early phase considerations, it would be convenient if extremes of a reasonable accuracy could be obtained based on shorter and fewer simulations. It is therefore of interest to develop specific methods which make it possible to extract the necessary information from relatively short time histories. The method proposed in this paper opens up the possibility to predict simply and efficiently long-term extreme response statistics, which is an important issue for the design of offshore structures. A short description of this is given, but in the present paper the emphasis is on short-term analyses. The results presented are based on extensive simulation results for the Kvitebjo̸rn jacket structure, in operation on the Norwegian Continental Shelf. Specifically, deck response time histories for different sea states simulated from a MDOF model were used as the basis for our analyses.

On Efficient Long-Term Extreme Response Estimation for a Moored Floating Structure

2018 ◽  
Author(s):  
HyeongUk Lim ◽  
Lance Manuel ◽  
Ying Min Low
Keyword(s):  
Wave Train ◽  
Offshore Structures ◽  
Offshore Structure ◽  
Floating Structure ◽  
Sea State ◽  
Peak Period ◽  
Frequency Wave ◽  
Extreme Response ◽  
Term Response

This study investigates the use of efficient surrogate model development with the help of polynomial chaos expansion (PCE) for the prediction of the long-term extreme surge motion of a simple moored offshore structure. The structure is subjected to first-order and second-order (difference-frequency) wave loading. Uncertainty in the long-term response results from the contrasting sea state conditions, characterized by significant wave height, Hs, and spectral peak period, Tp, and their relative likelihood of occurrence; these two variables are explicitly included in the PCE-based uncertainty quantification (UQ). In a given sea state, however, response simulations must be run for any sampled Hs and Tp; in such simulations, typically, a set of random phases (and deterministic amplitudes) define a wave train consistent with the defined sea state. These random phases for all the frequency components in the wave train introduce additional uncertainty in the simulated waves and in the response. The UQ framework treats these two sources of uncertainty — from Hs and Tp on the one hand, and the phase vector on the other — in a nested manner that is shown to efficiently yield long-term surge motion extreme predictions consistent with more expensive Monte Carlo simulations, which serve as the truth system. Success with the method suggests that similar inexpensive surrogate models may be developed for assessing the long-term response of various offshore structures.

Long-Term Extreme Response Prediction of Mooring Lines Using Subset Simulation

2018 ◽  
Author(s):  
Darrell Leong ◽  
Ying Min Low ◽  
Youngkook Kim
Keyword(s):  
Monte Carlo ◽  
Response Prediction ◽  
Offshore Structures ◽  
Lead Times ◽  
Subset Simulation ◽  
Mooring Lines ◽  
Promising Alternative ◽  
Probabilistic Description ◽  
Extreme Response

Rigorous methods of probabilistic evaluations on long-term extremes are integral components in reliability research of offshore structures against overload events. Assessment across all conceivable sea states requires accounting for variabilities of long-term environmental loads and short-term stochastics, traditionally captured through extensive sampling or numerical expectation integration. The amount of environmental load variables render numerical integrations across high dimensions computationally prohibitive, while industry requirements of high return periods demand large Monte Carlo samples of timedomain dynamic analyses. Subset simulation offers a promising alternative to classic methods of statistical analysis, dividing ultra-low probability problems into subsets of intermediate probabilities. The methodology is uniquely advantageous for the assessment of heavy-tail overload events, which are unpredictably severe and occur at exceedingly rare frequencies. Subset simulation is experimented on a mooring case study situated in the hurricane-prone Gulf of Mexico, with the structure exposed to a joint-probabilistic description of wave, wind and current loads. The devised methodology is found to successfully evaluate hurricane-stimulated extreme events at ultra-low probabilities, beyond the feasible reach of Monte Carlo simulation at reasonable lead times.

On Efficient Surrogate Model Development for the Prediction of the Long-Term Extreme Response of a Moored Floating Structure

2020 ◽  
Vol 143 (1) ◽  
Author(s):  
HyeongUk Lim ◽  
Lance Manuel ◽  
Ying Min Low
Keyword(s):  
Wave Train ◽  
Offshore Structures ◽  
Surrogate Models ◽  
Offshore Structure ◽  
Sea State ◽  
Peak Period ◽  
Frequency Wave ◽  
Extreme Response ◽  
Term Response

Abstract This study focuses on the development of efficient surrogate models by polynomial chaos expansion (PCE) for the prediction of the long-term extreme surge motion of a moored floating offshore structure. The structure is subjected to first-order and second-order (difference-frequency) wave loading. Uncertainty in the long-term response arises from contrasting sea state conditions, characterized by significant wave height, Hs, and spectral peak period, Tp, and their relative likelihood of occurrence; these two variables are explicitly included in the PCE-based uncertainty quantification (UQ). In a given sea state, however, response simulations must be run for the associated Hs and Tp; in such simulations, typically, a set of random amplitudes and phases define an irregular wave train consistent with that sea state. These random amplitudes and phases for all the frequency components in the wave train introduce additional uncertainty in the simulated waves and in the response. The UQ framework treats these two sources of uncertainty—from Hs and Tp on the one hand, and the amplitude and phase vectors on the other—in a manner that efficiently yields long-term surge motion extreme predictions consistent with more expensive Monte Carlo simulations (MCS) that serve as the “truth” system. To reduce uncertainty in response extremes that result from sea states with a low likelihood of occurrence, importance sampling is employed with both MCS- and PCE-based extreme response predictions. Satisfactory performance with such efficient surrogate models can help in assessing the long-term response of various offshore structures.

A Discussion of Long Term Response Versus Mean Maximum Response of the Selected Design Sea State

2007 ◽  
Author(s):  
S. Haver
Keyword(s):  
Offshore Structures ◽  
Wave Period ◽  
Sea State ◽  
Safety Level ◽  
Characteristic Response ◽  
Annual Probability ◽  
Extreme Response ◽  
Quadratic Response ◽  
Term Response

The overall aim of the design process is to ensure that a structure can withstand with sufficient margin all foreseen load events. Offshore rules and regulations will typically define a set of rules to be followed by the designer. By fulfilling these it is tacitly assumed that the aimed safety level is achieved. The governing regulations and rules and, not the least, their interpretation may vary between various offshore regions. This may be the case both for the load side of the problem and the strength side of the problem. In connection with design of offshore structures, the largest variability and uncertainty is typically associated with the load side of the problem. The purpose of this paper is to discuss the consequences of two rather common definitions regarding the characteristic environmental load to be used for design purposes: i) The characteristic response is taken as the load for which the annual probability of exceeding it is q (here referred to as the q-probability load), ii) The characteristic response is taken as the mean extreme response of a sea state for which the annual probability of exceeding it is q (q-probability sea state). Results are shown for some generic response cases varying from a linear response case being rather insensitive to wave period to a quadratic response problem associated with a critical wave period band.

Reliability of TLP Tendons Under Storm Sea States

2005 ◽  
Author(s):  
Federico Barranco Cicilia ◽  
Edison Castro Prates de Lima ◽  
Lui´s Volnei Sudati Sagrilo
Keyword(s):  
Reliability Analysis ◽  
Limit State ◽  
Nonlinear Dynamic Analysis ◽  
Analytical Models ◽  
Cumulative Probability ◽  
Wave Forces ◽  
Ultimate Limit State ◽  
Sea State ◽  
Extreme Response

This paper presents a methodology for reliability analysis of Tension Leg Platform (TLP) tendons subjected to extraordinary sea state conditions like hurricanes or winter storms. A coupled approach in time domain is used to carry out TLP random nonlinear dynamic analysis including wind, current and first and second order wave forces. The tendons Ultimate Limit State (ULS) condition is evaluated by an Interaction Ratio (IR) taking into account dynamic combination among tension, bending and hydrostatic pressure. Expected long-term extreme IR is obtained through the integration of cumulative probability functions (CPFs) fitted to response maxima associated to individual short term sea states. The reliability analysis is performed using a time-integrated scheme including uncertainties in loads, tendon strength, and analytical models. Failure probabilities for the most loaded tendon of a TLP in Campeche Bay, Mexico, considering a 100-yr design sea state and the 100-yr extreme response generated by long-term observed storms are compared.

A Bayesian Machine Learning Approach for Efficient Integrity Management of Steel Lazy Wave Risers

2020 ◽  
Author(s):  
Rasoul Hejazi ◽  
Andrew Grime ◽  
Mark Randolph ◽  
Mike Efthymiou
Keyword(s):  
Machine Learning ◽  
Fatigue Failure ◽  
Learning Process ◽  
Predictive Models ◽  
Structural Integrity ◽  
Degradation Mechanism ◽  
Data Driven ◽  
Practical Implementation ◽  
Integrity Management

Abstract In-service integrity management (IM) of steel lazy wave risers (SLWRs) can benefit significantly from quantitative assessment of the overall risk of system failure as it can provide an effective tool for decision making. SLWRs are prone to fatigue failure within their touchdown zone (TDZ). This failure mode needs to be evaluated rigorously in riser IM processes because fatigue is an ongoing degradation mechanism threatening the structural integrity of risers throughout their service life. However, accurately evaluating the probability of fatigue failure for riser systems within a useful time frame is challenging due to the need to run a large number of nonlinear, dynamic numerical time domain simulations. Applying the Bayesian framework for machine learning, through the use of Gaussian Processes (GP) for regression, offers an attractive solution to overcome the burden of prohibitive simulation run times. GPs are stochastic, data-driven predictive models which incorporate the underlying physics of the problem in the learning process, and facilitate rapid probabilistic assessments with limited loss in accuracy. This paper proposes an efficient framework for practical implementation of a GP to create predictive models for the estimation of fatigue responses at SLWR hotspots. Such models are able to perform stochastic response prediction within a few milliseconds, thus enabling rapid prediction of the probability of SLWR fatigue failure. A realistic North West Shelf (NWS) case study is used to demonstrate the framework, comprising a 20” SLWR connected to a representative floating facility located in 950 m water depth. A full hindcast metocean dataset with associated statistical distributions are used for the riser long-term fatigue loading conditions. Numerical simulation and sampling techniques are adopted to generate a simulation-based dataset for training the data-driven model. In addition, a recently developed dimensionality reduction technique is employed to improve efficiency and reduce complexity of the learning process. The results show that the stochastic predictive models developed by the suggested framework can predict the long-term TDZ fatigue damage of SLWRs due to vessel motions with an acceptable level of accuracy for practical purposes.

A Monte Carlo Approach to Prediction of Extreme Response Statistics of Drag Dominated Offshore Structures

2007 ◽  
Author(s):  
A. Naess ◽  
O. Gaidai
Keyword(s):  
Monte Carlo ◽  
Weather Conditions ◽  
Offshore Structures ◽  
Drag Forces ◽  
Wave Elevation ◽  
The Monte Carlo Method ◽  
Strong Current ◽  
Highly Nonlinear ◽  
Extreme Response

The focus of the present paper is the extreme response statistics of drag dominated offshore structures subjected to harsh weather conditions. More specifically, severe sea states both with and without strong current are considered. The nature of the hydrodynamic forces acting on the structure becomes highly nonlinear. Additionally to the drag forces, the so called inundation effect due to the wave elevation, corrected to include second order waves, is also taken into account. In the present paper the Monte Carlo method along with a special extrapolation technique is applied. The proposed method opens up the possibility to predict simply and efficiently long-term extreme response statistics, which is an important issue for the offshore structures design.

Estimation of Extreme Ship Response

2012 ◽  
Vol 56 (01) ◽  
pp. 23-34
Author(s):  
Wengang Mao ◽  
Igor Rychlik
Keyword(s):  
Gaussian Model ◽  
Container Ship ◽  
Short Term ◽  
Sea State ◽  
High Quantiles ◽  
Alternative Approach ◽  
Extreme Response ◽  
Heading Angle ◽  
Ship Speed

In practice the severity of ship response is measured by high quantiles of long-term distribution of the response. The distribution is estimated by combining the short-term distribution of the response with a long-term probability distribution of encountered sea states. The paper describes an alternative approach, the so-called Rice's method, based on estimation of expected number of upcrossings of high levels by stress during 1 year. The method requires description of long-term variability of the standard deviation, skewness, kurtosis, and zero upcrossing frequency of ship response. It is assumed that the parameters are functions of encountered significant wave height, heading angle, and ship speed. The relation can be estimated from the measured stresses or computed by dedicated software assuming rigid ship hull model. Then Winterstein's transformed Gaussian model is used to estimate the upcrossing rates of response during a sea state. The proposed method is validated using the full-scale measurements of a 2,800 TEU container ship during the first 6 months of 2008. Numerical estimation of 4,400 TEU container ship extreme of the extreme response for a 4400 TEU container ship illustrates the approach when no measurements are available.

Sign in / Sign up

Export Citation Format

Share Document