Introduction of Hierarchical Bayes Model for the Evaluation of Pipe Failure Rate Using Cross-Industry Database of Pipe Failure
-
- OKAJIMA Satoshi
- Department of Mechanical Engineering, The University of Tokyo
-
- IZUMI Satoshi
- Department of Mechanical Engineering, The University of Tokyo
-
- SAKAI Shinsuke
- Department of Mechanical Engineering, The University of Tokyo
Bibliographic Information
- Other Title
-
- 配管破損率を産業横断型配管損傷データベースから推定するための階層ベイズモデルの導入
- ハイカン ハソンリツ オ サンギョウ オウダンガタ ハイカン ソンショウ データベース カラ スイテイ スル タメノ カイソウ ベイズ モデル ノ ドウニュウ
Search this article
Abstract
Building a cross-industry database of pipe failure is an effective method to improve the evaluation accuracy of pipe failure rate. Although pipes have many common failure mechanisms, they are used in many different environments in each industry. Thus, to enable practical use of a cross-industry database, it is desirable to develop a method for evaluating pipe failure rate that is capable of modeling the variability between different industries. The hierarchical Bayes model is one method that can model the variability between different data sources. In this paper, a method for evaluating the pipe failure rate for one industry is developed based on the hierarchical Bayes model. Even when the database contains small number of failure records, this proposed method is always capable of utilizing those records to evaluate the failure rate for a specific industry. Using the proposed method in conjunction with a cross-industry pipe-failure database, the pipe-failure rate was evaluated to a greater accuracy than that obtained using the conventional Bayes method.
Journal
-
- TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series A
-
TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series A 74 (741), 741-748, 2008
The Japan Society of Mechanical Engineers
- Tweet
Details 詳細情報について
-
- CRID
- 1390282679456432896
-
- NII Article ID
- 110006685202
-
- NII Book ID
- AN0018742X
-
- ISSN
- 18848338
- 03875008
-
- NDL BIB ID
- 9529465
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- Crossref
- CiNii Articles
-
- Abstract License Flag
- Disallowed