中文摘要：

能够反映实际环境荷载特征的概率模型是合理进行海洋平台结构设计和评定的基础。海洋平台设计环境荷载往往由沿岸台站资料直接得到或经相关分析推算得到,而反映当地环境条件的现场实测资料应用较少。发展了一种利用现场实测资料对环境荷载极值概率模型进行修正、从而获得融合现场资料信息和原极值资料信息的更新极值概率模型的计算方法。首先将经典Bayes方法进行推广到考虑多种先验信息来源的情况,建立了多种先验信息源下的参数估计公式;应用非正态随机变量的正态变换方法,结合海洋环境要素极值概率模型参数估计方法,将Bayes方法推广到极值Ⅰ型随机变量的情况,建立了海洋环境要素极值更新概率模型及参数估计方法。通过模拟算例和实际风速数据验证了海洋环境要素极值更新概率模型的合理性。

英文摘要：

A model that sufficiently describes the probabilistic characteristics of environmental loads is a prerequisite for rational load calibration for design and evaluation of offshore structures. The design loads for platform structures in the sea are usually estimated directly from the onshore records or using correlation analysis, while field records of environmental factors are not fully used. A method is developed to update the conventional probabilistic model of extreme values with the field records. The classical Bayesian method is extended for parameter estimation for cases when multi-prior-source data are available. By transforming the non-Gaussian random variables into equivalent Gaussian ones, the Bayes-based updating theory is extended to cases when variables are of extreme values of typeⅠ. The model is validated through examples using Monte-Carlo simulations and observed wind data.