報告書番号

S91002

タイトル（和文）

原子力発電所における運転員のチーム行動シミュレーション技術の開発（その２）

タイトル（英文）

Development of simulation Methods for the team Behavior in a Nuclear Power Plant Control Room

概要 （図表や脚注は「報告書全文」に掲載しております）

本報告では、人間の認知・判断過程の内外での研究状況ならびにそのシミュレーションに有効と思われるファジー理論、ニューラルネットワークなど、最新の計算機利用技術を調査した。また、具体例として認知・判断の場面が多く、プラントの運転と類似性が認められる車のドライバーを対象として認知・判断・行動過程をモデル化すると共に、そのシミュレーションシステムを試作し、システム化する際の技術問題点について検討した。その結果、仮説、検証時のエラーの必要性や短期記憶の重要性など認知・判断・行動過程をモデル化する上で留意すべき事柄ならびにそれをルール処理することの限界、ファジー理論、ニューラルネットワークを適用することの利点などについて明らかにした。

概要 （英文）

In order to discuss the human factors related problems of anticipated troubles in nuclear power plants, it is important to analyse the intention formation process untill concrete actions are taken. For this purpose, it is desired to discuss both 1 the communication and role assignment as a team and 2 how each operator recognizes and judges theenvironmental information as a person. This paper describes the study status of human cognition, decision and behavior and also demonstrates the simulation systemof a car driver. Technical problems when constructing the human cognition, decision and behavior process are discussed through these studies. The main results are as follows.Cognition, decision and behavior model of a car driver was made refering to themodel of Baron and Wickens. This model includes a decision bypass line and is also considered the effects of human character and stress. The following insights were obtained through the discussion of the model and the construction of the simulation system.(1) Thknowledges in the long term memory are directly called to the decision part in this system, but when considering the errors concerned to cognition and decision, some devices must be added to the system so that no knowledge is called to the partdirectly.(2) The system includes only the long term memory for the simplicity, but the short term memory must be introduced for more realistic system in the future.(3) The simulation system represents the cognition, decision and behavior process as a rule and adopts the fuzzy theory in the cognition, decision part. The theory is suitable for the formulation of a car driver's decision. (4)The neural networks which model the human brain cell seems to be difficult to be applied to such a system. But if the sensuous values of the distance and the speed when passing by the obstacle are put on the networks beforehand, the speed and the distance becomes more moderate than only the fuzzy theory.(5) To make the performance of a task smooth when another task isbeing performed, the system should be better to include a drive history and a knowledge to understand the meaning of the task.

報告者

キーワード