The advancement in technology, automation and mechanization has revolutionized the industrial sector and as well as the jobs of concerned operators. Process industry (including the oil and gas sector) are saturated with complex equipment, sensitive parameters, variables, precise flow rates and given set of conditions; fulfillment of those are necessary for smooth, continuous, profitable and most importantly hazard free processes. A minor error or misunderstanding of the operator can put the whole process at higher levels of risk. This paper presents a methodology for the anticipation and possible reduction to the risk exposure. A replica of an industrial plant is developed within a virtual environment facilitated by 3D glasses for stereoscopic vision and 3D spatialized audio for higher immersivity. This arrangement is coupled with a process and accident simulator. A detailed and immersive 3D model of the plant allows the operators understanding the details of both equipment and operating conditions. Moreover, the added augmented virtual reality feature allows enhancing and improving the understanding and skill of operators by letting them know the process operating conditions that dynamically change with the plant section where they work. Various abnormal scenarios of industrial processes can be developed by means of this the virtual environment tool. The operators will be exposed to these scenarios in a virtual environment in order to observe and evaluate their reaction and handling of situation for the anticipation of risk and its possible reduction. A set of experiments are performed in a dedicated 3D demo room where the subjects to be trained and assessed are either bachelor or master degree students. They are given adequate training before running the experiments. The results produced by the training and assessing procedure allow validating the proposed methodology and opening new horizons for the risk anticipation as well training procedures for risk reduction. The authors are involved in developing both methods and algorithms to exploit the operator assessment in virtual environments for decision makers and also for the sake of recruitment of operator.