Parallel co-simulation of heavy-haul train braking dynamics with strong nonlinearities

A new wave of research interests on train braking dynamics has been raised which requires more accurate and faster train braking dynamics simulations. Simulations with fully nonlinear fluid dynamic air brake models and nonlinear friction draft gear models are more accurate, however these are not yet available. This paper developed a model that has combined both of these aspects. Simulations were achieved by using a parallel co-simulation method which requires minimum working loads to merge legacy codes. A case study was carried out to simulate a heavy haul train braking commencing with a minimum service brake and then followed by a full-service brake before the minimum service brake was fully executed. Such transient simulations were not able to be accurately conducted by using empirical air brake models, but only by using fluid dynamic air brake models. Influences of co-simulation synchronization frequencies on computing time and results of the simulations were also investigated. Simulation results have recommended a synchronization frequency of 100 Hz. The corresponding simulation was 43% faster than real-time for a 152-vehicle train.