Assessment of the reliability of existing or proposed escape, evacuation, and rescue (EER) systems is a vital part of safety management for existing or new offshore installations. This paper will review the fundamental concepts of EER, present new methodologies for both deterministic (expected and/or worst case) and simulation modeling, and present applications of these models to frontier offshore installations. Case studies for open water jacket type production platform operations will be based on those carried out for the Sable Offshore Energy Project (SOEP) off the East Coast of Canada, and will include both deterministic and Monte Carlo simulation results. Two principal evacuation systems are used by SOEP and modeled here; namely davit launched TEMPSC and the Skyscape systems. Ways of probabilistically incorporating the interactive effects on the EER process of the initiating accident, seastate and weather, and availability of different rescue modes (standby or passing vessel, helicopter, land, or other platform) are presented and incorporated in the models. Although similar modeling approaches to installations in ice populated waters are used, special technologies relating to the evacuation processes to the ice surface or broken ice zone and ice capable safety craft had to be developed and introduced to complete the reliability evaluation for arctic offshore EER. Also, due to the integral dependence of the safety craft integrity and evacuee survival on ice conditions, a probabilistic procedure realistically yet efficiently simulating the ice conditions was developed and implemented in the model. Finally, for both open and ice covered water installations, integrated expected case and Monte Carlo simulation results are presented and discussed in terms of safety implications and developmental requirements. Conclusions and recommendations for future work are given.
Reliable Escape, Evacuation, and Rescue (EER) could have averted or reduced the catastrophic casualty consequences of marine disasters such as the Alexander Kielland, Ocean Ranger, or the Piper Alpha. This statement automatically gives rise to two questions. What is reliable? Could reliable EER really have helped? The initiating events for either of the above disasters were neither unexpected nor unpredictable, although they were serious. The Piper Alpha marine disaster  was initiated by a relatively small maintenance related gas leak, which rapidly escalated to encompass the entire installation; emergency procedures are well established for maintenance activities. In the case of the Ocean Ranger , a severe storm caused unexpected loss of ballast system control, which escalated to a loss of stability and relatively rapid catastrophic sinking. Again, the design limits of the structure were not exceeded in the environmental conditions that initiated the disaster.
So how could one have predicted what is applicable and successful EER process in either of the two cases? Undoubtedly, both installations had well established emergency response plans and conducted drills, including unannounced (surprise) escape and evacuation drills, on a regular basis. Unfortunately, no matter how realistic drills under non-emergency conditions are, they fail to simulate a real accident situation.