基于模糊聚类方法的海工结构耐久性评估

Durability assessment of marine structures based on fuzzy clustering method

考虑海工结构耐久性评估的模糊性特点,提出一种新的耐久性评估方法.基于模糊数学原理对待估样本进行聚类分析,利用MATLAB模糊工具箱实现对海工构件耐久性的模糊评估;建立样本对于类别的不确定性描述,给出每个样本隶属于某一聚类中心的模糊隶属度.结果表明:采用的模糊聚类方法避免了传统硬划分对构件耐久性评估的不合理性,符合客观实际;基于软划分的模糊c-均值聚类的MATLAB编程法避免了主观因素对评估结果的影响,可对多个样本及多个评价指标的待评估事物进行模糊聚类分析,评估结果形象且动态反映了样本的耐久性状况.

Taking the fuzzy character of the durability assessment of marine structure into account,a new type of durability assessment method was proposed.The samples were analysed for clustering based with the principle of fuzzy mathematics,and the durability of marine structures was assessed by using fuzzy toolbox of MATLAB.The uncertainty of the sample in connection with its classification was described,and the fuzzy membership of each sample,belonging to the clustering center,was given.The result showed that the new method avoided the unreasonable nature of traditional durability assessment due to its hard division of the components,being more objective.The MATLAB programming method,which is based on soft division and fuzzy c-mean clustering,avoided the impact of subjective factors on the result of assessment.The fuzzy cluster analysis could be done for multiple samples and multiple evaluation indicators by using this method.The result of assessment reflected the durability state of the samples dynamically and vividly.