Examples of application of Bayesian Probabilistic networks：贝叶斯概率网络的应用实例[精品].ppt

* * * John Dalsgaard S?rensen1,2, Jens N?rk?r S?rensen2 J?rgen Lemming2 1) Aalborg University, Denmark 2) DTU Wind Energy, Denmark Introduction Risk assessment - wind turbine parts Risk assessment - ice throw Acceptable risk level Conclusions Risk assessment of wind turbines close to highways Introduction Lystrup (Jutland, Denmark), February 2012 Scope: Assessment of the minimum distance from wind turbines to highways (in Denmark) based on risk assessments of the consequences due to total or partial failure of a wind turbine ice throw in case of over-icing Introduction Laakso et al. IEA 2009 Methodology: Data collected from a large number of modern wind turbines from Denmark and abroad same basic technology as new large wind turbines information on events where parts of the turbine is thrown / dropped at a distance from the turbine Risk estimated that persons in a car are killed because of wind turbine parts thrown away from a wind turbine in events with total or partial failure ice pieces thrown away from a wind turbine in events with over-icing Introduction Data from representative large databases with information on parts ’thrown’ in case of failure (from nacelle and blades): Distance from wind turbine Size of with turbine part Estimation of probability per year per m2 that a wind turbine part hits in radius R from the wind turbine Estimation of probability per km that a vehicle is hit by a wind turbine part thrown from a wind turbine placed in a distance d from a road – depending on Vehicle velocity V (80 km/h) Distance D between wind turbines Assumption: If a wind turbine part hits a vehicle then in average 1.5 persons are killed (conservative) Risk assessment – wind turbine parts General statistical risk to be killed when driving on a highway in Denmark: 2 10-9 per km driven (2009) ALARP (As Low As Reasonably Practicable) principle: An additional / extra risk