Operating Procedure Analysis with Consequence Prediction by Dynamic Simulation

Faults due to human errors cost the petrochemical industry billions of dollars every year. In this work, we propose an operating procedure analysis (OPA) to estimate the probabilities of faults due to human error while following standard operating procedures (SOP) such as maintenance procedures and equipment switching procedures. These procedures occur frequently during the operation of continuous chemical process plants. The proposed OPA is similar to event tree analysis (ETA) with branches and events, but has been specifically designed for analyzing operating procedures. Transitions involving more frequent or more complicated operator procedures are likely at increased risk of faults.

In this work we demonstrated the OPA methodology for both a high frequency daily dryer switching scenario and a relatively low frequency monthly reflux pump switching scenario. Both equipment switching scenarios are commonly accomplished manually by operators. First, SOPs were developed and potential operator errors were determined for each procedural step. Then, dynamic simulation was applied to determine the consequences of each error. Finally, the frequency of consequences resulting from those scenarios was calculated. An Aspen HYSYS dynamic simulation was developed to determine the specific consequence of each fault.

We found that the probability for a typical ethylene plant to flare during a dryer switching procedure is more than 90% per year. These results emphasize the importance of careful analysis of manual operating procedures for safe and environmentally friendly operation. Due to the high probability of flaring, the human factor and SOPs should be evaluated for changes that could reduce the procedure abnormalities. One alternative is to emphasize operator training and safety education; another is to redesign the procedure to reduce the probability of mistakes. Simplifying procedures would make them less susceptible to operator errors and decrease the probability of faults.