Introduction: To control a continuous production system whose components are exposed to failures, it is necessary to provide intelligence to the monitoring mechanism, due to the need to identify the type of failure, its source, and anticipate the consequences that arise from its occurrence. Methodology: In this paper, it is proposed to extend Sanz's multi-resolution model and apply it to a supervision scheme with event detection based on fuzzy logic and implemented using agent technology. Results: A mechanism to validate the implementation using discrete event simulation is also presented. Conclusions: It was concluded that discrete event simulation constitutes an appropriate way to validate a supervisory control system at a high level.