Introducing the Rail Accident Analysis Model and Algorithm by Using Neuro-Fuzzy Intelligent Systems (Case Study Derailment)

Due to the safety of Rail transportation, Investigation on characteristics and phenomenon which causes accidents has an important role, in order to enhance the safety of rail transport and, in turn, reduce the rate of accidents. According to the rail accident statistics, among all factors causing train accidents, derailment has the biggest share, which can be because of human errors, track and rolling stock deficiencies, etc. Investigation and technical analyses of these factors will be complicated, because of the quantitative/qualitative relationships. Investigation on available derailment results indicates that a few numbers of traditional analytical models have been used. Herein, limited capability of these models such as being one dimensional, taking not into account all factors involved and complexity in behavior description can be clearly observed. Regarding the importance of the current situation in the derailment investigation, in this research a new combined intelligent Neuro-Fuzzy model has been developed. This study has simplified the quantitative/qualitative behavior description for derailment, prepared and simplified a practical and more accurate quantitative/ qualitative evaluation to achieve multi-dimensional decision making evaluation model, by using Adaptive Neuro Fuzzy Inference System (ANFIS). This research has developed a method of operation based on human factors. Regarding the results, analyses of the developed N-F model make a foundation for a more comprehensive accident management system.