報告書番号

N07027

タイトル（和文）

地震後の災害情報を逐次処理する配電設備被害推定の基本モデル－ベイジアンネットワークを適用した被害推定システムの開発－

タイトル（英文）

Basic model for the damage estimation of electric power distribution facilities based on a sequential processing of the disaster information after an earthquake-Development of a damage estimation system using Bayesian network-

概要 （図表や脚注は「報告書全文」に掲載しております）

本報は，配電施設を対象として，地震直後から応急復旧完了までの情報錯綜期を主に支援する逐次更新型被害想定システムの基本モデルを提案することを目的とする．
面的に膨大な数存在する配電施設は大規模地震時に同時多発的に被害が発生するため，効率のよい事前対策・復旧作業を支援する高精度な被害想定を実現することが要求されている．しかしながら，配電施設の被害想定を精度良く実施する場合には，地震動強度，地盤条件，自社設備の力学的な特性，および樹木や建物など周辺施設の力学的な特性まで考慮したモデル化が必要となり，一般に，すべての情報を事前に把握し，被害想定を実施することは困難である．
これに対し，提案システムは，ベイジアンネットワークを応用し，配電施設被害に関連する災害事象を逐次更新情報として収集し，配電施設被害想定精度を向上させることが可能である．本報では，地震動強度情報と停電情報の処理プロセスを定式化し，既往被害想定と比べ，特に，地震直後の配電施設被害想定を高精度に行うことが可能となることを例示した．

概要 （英文）

This report presents a basic model for a seismic damage estimation system based on sequential data processing of multi-disaster information to support its emergency restoration activity. Since a huge number of electric power distribution facilities are spatially installed in local area, a lot of facilities are simultaneously damaged in a semsic event. Thus, a highly accurate damage estimation system for the distribution facilities is needed to realize their effective seismic countermeasure and restoration. However, in order to accurately estimate the seismic damage of the distribution facilities, it is needed to collect not only the detail information concerning a dynamic characteristic of its own equipment, but also those of seismic intensity, ground condition, and dynamic characteristics of local facilities including trees and general frame houses. It is usually difficult to collect such detail information. The proposed system can improve the accuracy of the damage estimation based on some sequential multi-disaster information including the information of the power outage after an earthquake event. Its model is based on the Bayesian network modeling process. In this report, the damage estimation process is formulated based on the two post-earthquake information including seismic intensity and power outage, which can be collected just after an earthquake event. In numerical examples, it is illustrated that the proposed model enables us to more accurately estimate the damage of the distribution facilities just after the earthquake event compared with the existing damage estimation models.

報告者

キーワード