Abstract: To reveal the law of risk propagation and to identify the root cause of schedule uncertainty, a method of engineering schedule risk analysis is developed with the risk-driven principle as the basic idea, based on the risk chain theory and combined with the Bayesian Belief Networks and the Monte Carlo technology to qualitatively-quantitatively describe the risk evolution process. In this method, the correlation of risk factors, the multiple uncertainty states of risks, and the rationality setting of the simulation environment are considered. Engineering application shows that its maximum error of the simulated average total construction period is only 1.5%, compared with the Monte Carlo simulation method and the standard correlated schedule risk analysis model (CSRAM). Its sensitivity analysis function can be used to mine the key risk triggers at the project level and each activity level, a useful device for managers to formulate relief countermeasures scientifically.

Key words: project progress management, risk chain, schedule risk analysis, risk-driven, multiple uncertainty