This paper proposes a Bayesian-theory-based risk analysis method of coastal structure design criteria along Qingdao coasts of Yellow Sea. New method has some advantages over traditional method. This is because Bayesian theory continually involves new data and information in the process of prior probability and posterior probability exchanges and then improves the model of statistical prediction. The proposed method gives quantitative risk analysis of coastal structural stability.


Traditional design method of coastal structures always bases on the calculation by design codes or experimental modeling tests. And sea environmental events are chosen as determinate combinations as input parameters. Different kinds of uncertainties of sea environments and sensitivity of structural responses to uncertainties of input parameters are not taken into consideration. However, uncertainties of wave heights, wave periods and water levels are remarkably important for coastal structures. The Bayesian theory based analysis method through the circle of prior probabilities--posterior probabilities involves different kinds of uncertainties and their corresponding correction during Bayesian analysis processes (Box and Tiao, 1973; Neter, Wasserman and Whitmore, 1988). Finally the most disadvantageous response for structural stability can be selected.


For following extreme sea environments induced wave forces and structural stability analysis, after first calibration circle of probability distribution models, for example, structural response induced by extreme wave height, the outcome state probability can be gotten. It is also called as prior probability because the sample information subsequently provides further information about the prevailing outcome state. Next, for the prior probability and the new data series (including different kinds of uncertainties) the posterior probability () i P M (i) P Mi x () i i P M x can be obtained using Bayesian t y.

This content is only available via PDF.
You can access this article if you purchase or spend a download.