This Small Business Innovation Research Phase I project from Gensym Corporation will investigate the feasibility of developing online, real time reliability and risk analysis software that will analyze process data to ascertain the reliability of equipment and system functions, assess the current risk in the facility, report and possibly alarm operators and managers of the risk to system functions and equipment, and recommend corrective actions to maintain control over risk. For many organization, expenditure on risk mitigation and safety is significant. Risk mitigation costs can be reduced by more carefully monitoring risk and taking corrective action before worst case scenarios develop. The greater the ability to monitor and the tighter the feedback loop between risk discovery and evasive action, the more significant the savings. This research will demonstrate that Bayesian networks provide a core representation for most reliability and risk analysis models, that system decomposability can be exploited to provide real time assessment, and that learning methods can be utilized to estimate and recalibrate the analysis. The value for an online risk management capability is high in many industries: Environmental monitoring, toxic waste processing, cleanup and shipment, nuclear power generation, offshore oil production and shipment, emergency response and emergency evacuation, critical telecommunications, transportation, flexible manufacturing, securities trading, credit fraud prevention, and many others.