ABSTRACT
Accident prevention is a central issue in high-risk industries such as nuclear power plants, aircrafts and petrochemical plants. These industries have set up safer equipments, technical safety barriers but also organizational and human barriers in order to manage risk and improve their capacity to prevent the accidents. Some researches admit that industries’ capacities to learn from their own experiences will enable them to improve accident prevention. Reporting systems such as Learning fromExperience (Retour d’Expérience in French) aim at learning from failures and negative experiences. They are implanted to collect, analyze and share data on the accidents which occurred on sites. Despite the high relevance of Rex System, many researchers (Bourrier, 2002, Dien, 2006 and Amalberti and Barriquault, 1999) showed two main weaknesses: the limits-Rex would provide mainly technical and direct causes of the accidents, and biases-Rex would be more used as an enormous data base than an opportunity to share the lessons learnt from the accidents. The goal of this research is trying to overcome these limits by exploring new research areas. The issue of weak signals is emerging in industrial companies like EDF (Electricité de France1) and academic researches might provide relevant ideas. Defined as accidents precursors, the weak signals would enable to identify unsafe situations and degradation of the system. In that respect, identifying
and taking the weak signals into account would favor proactive approaches and better accidents prevention.
