事故調查及成災因素分析應用於風險評估和災害預防之整合研究-總計畫暨子計畫:重大災害事故調查機制與程序之研究

Abstract

在台灣工業災害歷程上，每當發生了對社會產生深刻影響的重大災害時，企業才會開始 從中學習，把災害造成的教訓及經驗應用到企業內部各個方面。而由工業災害來看，不 論是高科技產業或是傳統產業，同樣類型的災害事故一再發生的情況是很常見的，究其 原因之一，就是對災害發生原因的鑑識及探討機制不完備。故此，本計畫以工業災害事 故為主，以三年時間，透過系統化分析技術進行事故原因的分析鑑認，藉以對操作及管 理系統缺陷進行鑑識及提出改善，以達到杜絕工業事故再發生為主要目的。本三年計畫 內容簡述如下：第一年首先進行基本資料收集，進行系統化歸納及分析方法分類，然後 發展建立事故模式(包括有連續事件模式，類流行病模式及系統事故模式分析)，以評估 事故模式發展可能性與有效性，最後發展事故調查流程，建立證據收集程序及分析程序 來以供改善建議。第二年主要是建立事故調查方法及程序，首先針對個別方法分別進行 深入的探討，這些方法包括有屏障分析、變更分析、事件及成因圖、意外事件發生原因 分析、系統原因分析技術、連續時間事件圖、失誤樹分析、管理疏忽與風險樹、事件樹 分析、為何樹分析、原因樹分析和作業事故調查技術，然後再進行整合來完成及發展最 適化調查方法。第三年則依據第二年所建立之最適化事故調查方法來發展工具，他們包 括有事故背景紀錄表單，時間序列圖，查核表，數據資料庫，然後建立原因因子探討分 析技術，這些技術包括有直接原因、間接原因及根本原因分析技術，最後統籌作案例分 析及應用。

From the observation of severe industrial incident history in Taiwan, the industrial sector only learnt the bitter lesson and felt the great pain as the tremendous influential impact on the society took place due to the occurrence of hazardous accident. Then, the owners applied such lesson and experience to the management. From the viewpoint of industrial hazards, it is very common to see that the similar incidents happened again and again no matter what high-tech or traditional industrial. One of the basic reasons is that the scene investigation and corresponding mechanism for root causes of incident are not complete. Therefore, this three-year project focuses on the studies of industrial incidents to develop and formulate a systematic analysis technique to carry out the incident investigation to find out the root cause(s) for such matter. Then, correct or improve the defects of operation processes and management systems via PDCA by using this systematic analysis technique to achieve the goal of refusing the re-occurrence of similar industrial incidents. The contents of the first year projects are to collect the relevant literatures and to analyze, summarize and categorize. Then, develop the establishment of incidental models, including sequential, epidemiological and systemic accident ones to evaluate possibility and effectiveness of these modes. Eventually, develop the incident investigation procedure and build up the data collection and analysis processes for further improvement. In the second year, the main task is to establish the incident investigation methodology and procedure. First, it is to carry out the detailed analysis for each method. These methods include BA, CA, ECFC, TRIPOD, SCAT, STEP, FTA, MORT, RTA, WTA, CTA and WAIT. Then, integrate these methods to complete and develop the optimal investigation methodology. In the third year, it will to develop the correspondent tools incorporating with the optimal investigation methodology, developed in the previous year. They consist of the background record list, time evolution chart, check list and data bank. Then, build up the cause factor analysis techniques, which include the direct, indirect and root cause analysis ones. Finally, several case studies will be considered and applied according to these techniques.