Risk management on offshore platforms requires real-time adaptation of procedures to existing conditions. These conditions can be caused either by environmental factors(e.g., storms or ice), on-board incidents(e.g., a fire), boat collisions, or degradation of the platform or its capacity(e.g., by corrosion.) In situations of risk, the operators generally follow procedures that have been predesigned, but some specific circumstances may not have been foreseen in the design of the procedures. An experienced operator can generally estimate the severity of a situation similar to one he has experienced previously and take appropriate measures. In some instances, however, more information is needed about the current risk situation, potential deterioration of this condition, applicable procedures, and alternative options before appropriate actions can be taken. The object of this research is to combine the powers of analytical methods (probabilistic risk assessment based on systems analysis) and artificial intelligence (expert systems based on recommendations of experienced operators and on operating/emergency procedures) in a hybrid real-time decision support system for risk management on board platforms. A key theoretical issue is the aggregation of these two types of information. A prototype will be developed. One possible option for this prototype is to address the problem of safety management for a redundant platform when the weather conditions make both delayed and immediate maintenance hazardous.
