fa Railway track and structures Railway track and structures پژوهشي Research Safety is one of the superiorities of the railways in comparison to other modes of the transportation. Assessing the level of safety (LOS) in a railway network requires the up-to-date risk model which considers the hazards participating in an accident while analyzing the severity of their consequences. Bayesian networks analyze the conditional probabilities of the chain causes and demonstrate it as a visual graphical model for the relations between the involved factors. The accuracy of calculations depend on the recorded data and the main contributor of the model validity. In analyzing the risk of derailment on switches as one of the most recurrent accidents, different reasons which separately or combined with other factors are evaluated and categorized in six different severity categories based on the Iranian railway accident severity guidelines. The outcome of the model shows the human factors and its subdivisions as the main factor participants in the switch derailment and the risk of the third grade of the severity as the most probable severity degree. Railway safety assessment, Bayesian network, Risk model, Derailment, Model validation 103 112 http://ijrare.iust.ac.ir/browse.php?a_code=A-10-79-2&slc_lang=fa&sid=1 Shahin Shabani 180031947532846002087 180031947532846002087 Yes Department of Civil Engineering, Payame Noor University (PNU), Tehran, Iran Jabbar Ali Zakeri 180031947532846002088 180031947532846002088 No School of Railway Engineering, Iran University of Science and Technology, Tehran, Iran Vahid Rezaei Tabar 180031947532846002089 180031947532846002089 No Department of Civil Engineering, Payame Noor University (PNU), Tehran, Iran Azadeh Ghanbarpour 180031947532846002090 180031947532846002090 No Department of Civil Engineering, Payame Noor University (PNU), Tehran, Iran