3 results
Now showing1 - 3 of 3
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Time-variant rule-based reliability of corroded structures by Monte Carlo simulationGeneral widespread corrosion is of increasing concern for structural reliability of ageing marine structures, particularly for semi-submersible pontoons. Conditional failure rate, also termed hazard function is sought to describe the increasing maintenance required with age. Literature is reviewed connecting corrosion degradation to hazard rate, either through a time degrading limit state function, or statistical data for time to failure. Here, a rule based failure definition is applied based on corrosion tolerance levels, together with a linear corrosion degradation model to make a time variant limit state function. Monte Carlo simulation is applied based on statistical models for the basic variables including importance sampling, to obtain the time to failure distribution from which the hazard curve is deduced. A more realistic hazard function is also produced, requiring average of 3 simulations to exceed the tolerance in order to fail. It is seen that this approach delays the increase of the hazard curve, at the expense of higher failure rates.Publicationtype: Conference PaperScopus© Citations 5 334 213 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Power spectrum for surface description of corroded ship structure from laser scanStrength of corroded marine structures is in the industry traditionally assessed using average thickness reduction together with the assumption of a uniform surface. In academia it is however realised that the surface texture will influence the strength where the effect of the irregular surface is considered through various parameters such as reduced material properties, roughness, geometrical ratios and thickness distributions. In this paper, the objective is to present a power spectrum which describes the corroded surface topography. Therefore, power spectra of surface and thickness are produced from 3D laser scan of real corroded mild steel specimens from a 30 year old ship using Fast Fourier Transform (FFT). The frequency distribution describes the spatial variability, wherefrom the autocorrelation function (ACF) is deduced. A log-log fit is made to the power spectrum, where the interception with the ordinate axis describes the corrosion severity. This way of describing the surface pattern of the corrosion deterioration may serve as input to novel ways of calculating corroded strength and fatigue initiation as well as simulation of corroded surfaces.Publicationtype: Conference Paper322 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Updatable spatio-temporal probabilistic corrosion modeling for offshore structures(American Society of Mechanical Engineers (ASME), 2018-06); ; Corrosion models are important to assess how the corrosion influences current and future structural strength. For this purpose it is desirable to describe the uneven corrosion diminution of the irregular surface (i.e. Space) and progression (i.e. Time) in various corrosive environments. Thickness measurements give an indication of the current state, and should be considered in the corrosion model. The inherit uncertainty in corrosion argues for a probabilistic type of corrosion model. Probabilistic models to describe corrosion in time and space, and that can be updated with observations exist, but are typically too complicated for practical engineering use for in-service corrosion assessment. Simpler models exist, that do not describe all of the mentioned aspects (probabilistic, updatable, describe corrosion in time, space and various environments). Here, a simple model covering these aspects is described in two parts. First bayes updating is used to estimate the parameters of the corrosion distribution for each unique environment. The second part uses this resulting distribution and describes how this distribution develops with time. The model is demonstrated with an example and compared to similar spatio-temporal models. The model is promising for improvement from simplistic uniform description of surface and linear progression used in current industry practice.Publicationtype: Conference Paper140