The design of conventional safety systems is based on failure likelihood and accident severity, which is normally obtained empirically, leaving the system vulnerable to process nonlinearities. To ensure process safety, control actions are conservative and small deviations from setpoints may lead to shutdown, generating economic losses. In this work, periodic simulations of system behavior against failures is proposed in order to determine the potential risk to which the system is subjected. Depending on this potential, preventive actions can be taken in order to guarantee the system safety and integrity and avoid potential shutdown. These actions are calculated to provoke least possible disturbance in order to reduce impact on product quality, while keeping the process operating. The goal is to increase annual operating time of the plant without compromising safety and product quality. Results show that the proposal is feasible to real time applications and unnecessary shutdowns can be avoided. (C) 2016 Elsevier Ltd. All rights reserved.