報告書番号

S99009

タイトル（和文）

保守作業時のグループエラー行動予測に関するプロトタイプシステムの開発と性能評価

タイトル（英文）

Development and performance evaluation of a prototype system for prediction of the group error at the maintenance work

概要 （図表や脚注は「報告書全文」に掲載しております）

原子力発電所におけるより一層のエラーのゼロ化を図るため、保守作業時のグループエラー行動を予測するためのプロトタイプシステムを開発した。その特徴は次の通りである。(1)利用者が本システムに、予測対象としての「保守作業の特徴要因」、「組織・体制要因」および「グループPSF要因」の存在の有無と程度を入力することにより、7段階の階層からなる予測因果モデルを介して、目的とする保守グループエラーなどを予測できる (図1参照)(2)本システムは、「連携エラー」を最も高い割合で予測し、ついで「検知エラー」を高い割合で予測している。そして「意志決定エラー」、「伝達エラー」および「状態認知エラー」に対しては、ほぼ同じような割合で予測する特性を有している。(図2参照)

概要 （英文）

In order to attain zero-izing of much more error rather than it can set to a nuclear power plant, Authors development and its system-izing of the error prediction causal model which predicts group error action at the time of maintenance work were performed. This prototype system has the following feature. (1) When a user inputs the existence and the grade of the existence of the "feature factor of the maintenance work" as a prediction object, "an organization and an organization factor", and a "group PSF factor" into this system. The maintenance group error to target can be predicted through the prediction model which consists of a class of seven stages. (figure 1 reference) (2)This system By utilizing the information on a prediction result database, it can use not only for prediction of a maintenance group error but for various safe activity, such as KYT (dangerous forecast training) and TBM (Tool Box Meeting).(Figure 1 reference) (3)This system predicts a "cooperation error" at highest rate, and, subsequently predicts the "detection error" at a high rate.And to the "decision-making error", the "transfer error"and the "state cognitive error", it has the characteristic predicted at almost same rate.(Figure 2 references) (4)if it has full knowledge even of the features, such as the enforcement conditions of maintenance work, and organization, even if the user has neither the knowledge about a human factor, nor experience, anyone of this system is slight about the existence, its extent, etc. of generating of a maintenance group error made difficult from the former logically and systematically easily,it can predict in business time for about 15 minutes

報告者

キーワード