Virtual reality contribution to training and risk prevention

Industries need to train their employees continually to learn best practices but also to learn to react to disrupted situations, in particular at SEVESO sites. However, training in such sites is so limited due to potential risks produced by human errors and environmental conditions. In this context, Virtual Environments for Training/Learning (VET/L) are best suited to overcome these limits. In our work, we developed a VET/L equipped with an Intelligent Tutoring System (ITS) called HERA. In this paper, we present HERA that allows to track learners, detect their errors and prevent potential risks. In addition to classical ITS models, HERA contains a risks model to describe the causality relationships between risks and humane errors. These relationships are represented in a Bayesian network, which is built by interpreting the concepts of a risk analysis bow-tie. Moreover, we integrated a risks module that allows to evaluate risks severity.