Intelligent Railway Traffic Management智能铁路运输管理.pptx

Intelligent Railway Traffic Management;Content;Traffic management process;Monitoring system;Traffic control screens DB;Online train graphs Deutsche Bahn;Train connection supervision DB;Automatic conflict recognition/speed adaptation RCS SBB;Centralized Traffic Control;Current rescheduling practice;Current traffic management drawbacks;Essential requirements for railway perturbation management;Real-time traffic state monitoring and prediction;Three-layer process modelling;Train detection;Dutch process mining tool TROTS;Screen shot of TROTS logfile;Flow chart of process mining algorithm;;Blocking time;Automatic conflict detection tool (TU Delft) ;Automatic conflict detection tool (TU Delft): Blocking time diagram;(Re-)scheduling decision support models;Shortcomings of microscopic simulation models for real-time rescheduling;Intelligent Rescheduling (1);Intelligent Rescheduling (2);Intelligent Rescheduling (3);Feedforward traffic management information;Statistical learning methods;Dependence of running time on departure delay IC line 2200;Dependence of dwell time on arrival delay and time of the day in Delft;Current real-time information of train delays ;Running time prediction in case of delays;Monitoring and prediction in the traffic control loop;Based on historical data percentiles of process times are computed: Sum of running times over route segments (outbound route – open track blocks– inbound route) Dwell times Headway times between similar train movements at the point of conflict Connection times between same train series Running and dwell times are updated every 10 – 30 s based on current train positions and current delays Simple model can be extended by clustering historical data and classifying train runs according to: Time of day Rolling-stock type Weather Delay;Change of signal aspects and track occupation/ clearance times explain train movements;Validation of short-term prediction model;Example infrastructure elements of line Rotterdam Centre