Seismic Damage Prediction of Buried Pipeline by KDD Method
M.B. Javanbarg, S. Takad and Y. Kuwata
ABSTRACT: Damage prediction of buried pipeline under earthquake environments is the first stage for the seismic risk analysis. In this paper, we use a Knowledge Discovery in Database (KDD) method for the pipeline damage prediction even though many studies have been performed so far with the aid of empirical, statistical, and/or theoretical methods. By employing the KDD method, much higher accurate damage prediction could be done for better understanding of pipeline damage distribution. Related factors were analyzed by a GIS based model of the Kobe water buried pipelines in the 1995 Kobe Earthquake, and a decision tree of pipeline damage classification was developed based on the Classification and Regression Tree (CART) method. A verification of the method was focused to the modeled area, and accuracy of the proposed prediction method was confirmed in comparison with an actual damage as well as predicted ones by commonly used formula of damage estimation. Results of the developed KDD model showed that the model could predict correctly the number of damage in pipeline network. The proposed method by KDD turned out the distribution of damage better than other damage estimation methods.
Vol 8, No. 1 – Spring 2006
/در مجله JSEE /توسط adminSeismic Damage Prediction of Buried Pipeline by KDD Method
M.B. Javanbarg, S. Takad and Y. Kuwata
ABSTRACT: Damage prediction of buried pipeline under earthquake environments is the first stage for the seismic risk analysis. In this paper, we use a Knowledge Discovery in Database (KDD) method for the pipeline damage prediction even though many studies have been performed so far with the aid of empirical, statistical, and/or theoretical methods. By employing the KDD method, much higher accurate damage prediction could be done for better understanding of pipeline damage distribution. Related factors were analyzed by a GIS based model of the Kobe water buried pipelines in the 1995 Kobe Earthquake, and a decision tree of pipeline damage classification was developed based on the Classification and Regression Tree (CART) method. A verification of the method was focused to the modeled area, and accuracy of the proposed prediction method was confirmed in comparison with an actual damage as well as predicted ones by commonly used formula of damage estimation. Results of the developed KDD model showed that the model could predict correctly the number of damage in pipeline network. The proposed method by KDD turned out the distribution of damage better than other damage estimation methods.