Abstract: This article is a study about a knowledge-based complex system safety risk model. First, a know-ledge representation model of complex system safety occurrence cases is presented. Second, the uncertainty of knowledge representation as well as the requirement for uncertainty quantifications by the safety risk decision-maker is discussed. Thus, a knowledge-based system safety risk model is presented, where risk is a function of loss probability, loss degree, and loss uncertainty. Then, the modeling method of the proposed model is given, and an instance to demonstrate its application is provided. Finally, the model is used in the safety risk assessment of a civil aircraft flight operation. The attributes of pilot age, total pilot flight hours, and pilot flight hours in the past 90 days are selected as risk factors to assess the flight safety risk of the pilot. The assessment result demonstrates the validity of the proposed model.

Key words: safety, uncertainty, risk, risk-aversion, system, loss, knowledge representation