首页    期刊浏览 2024年11月25日 星期一
登录注册

文章基本信息

  • 标题:損傷情報のフィードバック手法に関する研究
  • 本地全文:下载
  • 作者:山本 規雄
  • 期刊名称:日本造船学会論文集
  • 印刷版ISSN:0514-8499
  • 电子版ISSN:1884-2070
  • 出版年度:1993
  • 卷号:1993
  • 期号:173
  • 页码:385-392
  • DOI:10.2534/jjasnaoe1968.1993.385
  • 出版社:The Japan Society of Naval Architects and Ocean Engineers
  • 摘要:

    In the hull structure design, it is necessary to consider various uncertainties due to the complexity of the structure, the severity of the environment, and the randomness of the wave induced loads, etc. Especially, in the design of new type of structure or peculiar structure which has not enough past experiences to use or when a rational design and a reliability based design are aimed, these consideration become more important. Moreover, it is important to take measures of the factors which has not taken into account in the design stage and which has not been able to evaluate enough in the design stage. For this purpose, the use of the obtained damage data is indispensable and it is necessary to feed back such data to the process of evaluation of safety and the reliability of the hull structure. In this paper, discussion is made on how to feed back the damage data to the design material or the repair material as the damage measures. The method based on Bayeian model is discussed as following processes by taking the fatigue damages caused in the hull structure into account as an example. (1) Formulate the mechanism of occurrence of fatigue damage with the condition that the state of the stress of each point is an unknown. And then, the likelihood function of these unknown parameters is composed when the data concerning the presence of finding the fatigue damage is obtained. (2) Introduce the probabilistic distribution explaining a restricted condition which have to fill by unknown parameters. And then the prior distribution, with the parameter which shows strength of the restricted condition as an unknown parameter, is composed. (3) Evaluate unknown parameters which explain the state of the stress of each point based on the posterior distribution derived from Bayes' theorem. (4) Evaluate the other unknown parameter of the likelihood function and the unknown parameter of the assumed prior distribution by applying the most likelihood estimation method. In this case, liklihood function is defined as the marginal distribution concerning the assumed prior distribution. This contains the problem hw to do the comparison of Bayeian models. In this case, the one that the AIC (An Information Criteria) expanded for Bayeian model is used. Moreover, in this paper, the state of the stress of each point is assessed based on the actual data concerning the presence of finding the damage reported in some ships considered to be same structural type. In addition, based on these results, an example of feeding back the damage data to the repair material as the damage measures is discussed.

国家哲学社会科学文献中心版权所有