中文摘要：

不安全动作的定义、分类与识别尚未有统一标准,这给事故分析与预防带来许多困难。基于此,对人因分析与分类系统（HFACS）和24 Model中的不安全动作层面进行对比研究。在不安全动作对应方面：HFACS中的不安全动作均能对应到24 Model中的不安全动作,但HFACS没有涉及24 Model中不违章、未引起事故但高风险的不安全动作,因而无法实现事前预防。在间接原因对应方面,HFACS中不安全动作的产生原因均可与24 Model中的1个或多个间接原因相对应;HFACS中不安全动作的产生涉及较多生理、心理层面的原因,而考虑到实用性原则,24 Model暂未给出该两方面原因的具体分类方法。在不安全动作发出者方面：24 Model涉及组织内各层级人员,而HFACS仅指一线员工。在实际应用方面：两种模型均可用于不安全动作原因的统计分析;24 Model给出的不安全动作分类范围较HFACS广,在事故分析时所受局限性小,但易造成不安全动作遗漏;在用24 Model进行不安全动作分析时,应尽量细化动作分类及分析的深入程度,以及动作发出者的员工层级。

英文摘要：

The present paper intends to propose a HFACS（Human Factors Analysis and Classification System） and a so-called＂24＂,or Accident-to-Causation Model（24 Model） involving both the relevant information and the cause over the unsafe acts.On the basis of the above-mentioned model,we have done a comparative study on the unsafe acts between the HFACS and 24 Model.According to the above 24 Model,it would be possible for all kinds of reckless or unsafe acts in HFACS to be attributed to one of the acts illustrated in the 24 Model,though the HFACS model does not include the high-risk unsafe acts that do not violate any of the regulations in operation or in production,which may account for why the HFACS model fails to do the prediction or forecast of the early accident prevention.As to the involvement of the corresponding indirect causes,what the HFACS model can do is just to contribute such unsafe acts to one or more indirect causes（knowledge,awareness,habit,psychology and physiology） in the model,for such unsafe acts involve more physiological and psychological factors.Nevertheless,it won＇t be possible for the 24 Model to give the specific taxonomy so far.Besides,the doers of the unsafe acts may include all staff-members at all levels（the frontline workers,the group leaders,the intermediate-level managers and even the senior-level managers,or known as CEO）in the organization or company systems whereas HFACS can only involve fore-front level employees,that is,the people who commit the unsafe acts in 24 Model.As for the practical applications of the 24 Model,the above mentioned two models can both be used to perform the statistics of the unsafe acts in majority of the accidents to obtain their happening frequency.Therefore,in comparison with the model of HFACS,the taxonomy of the unsafe acts of the 24 Model proves much wider and more complicated,which is not likely to limit the image or prediction of the accident analysts,but also is prone to result in the omission of unsafe acts.Furthermore,the 24 Model c